diff --git a/docs/Bharat_AI_SoC_2026_27.html b/docs/Bharat_AI_SoC_2026_27.html index 00339aa7..973b946a 100644 --- a/docs/Bharat_AI_SoC_2026_27.html +++ b/docs/Bharat_AI_SoC_2026_27.html @@ -1,7 +1,7 @@ --- layout: page title: Bharat AI-SoC Challenge — 2026–27 -published: false + show_title: false description: Bharat AI-SoC Challenge — 2026–27, 2nd Edition --- @@ -27,17 +27,21 @@

Bharat AI-SoC Challenge — 2026–27

Format - Three-Phase Innovation Pipeline + Open Qualifying & PoC Development
+
+ Register Now +
+

Overview

-

The inaugural Bharat AI-SoC Student Challenge attracted more than 3,000 students across 1,238 teams. Through three levels of evaluation and virtual mentorship, 150 shortlisted projects addressing five industry and societal problem statements were narrowed to five winners.

-

The 2nd Edition scales the challenge nationally, widens eligibility and introduces a multi-stage qualification process. Additional milestone-based grant support will help top teams progress beyond Proof of Concept (PoC) toward working prototypes and product readiness.

+

The inaugural Bharat AI-SoC Student Challenge, organised by Arm Education, MeitY C2S and IIT Delhi as Lead Mentor, attracted more than 3,000 students across 1,238 teams. Following three levels of evaluation and virtual mentorship, 150 shortlisted teams were narrowed to five winners.

+

The 2nd Edition expands eligibility nationally to students and working professionals / developers. Participants compete on equal terms across five problem statements, progressing from Open Qualifying to Proof of Concept (PoC) Development.

@@ -48,21 +52,21 @@

What’s New in the 2nd Edition

Students (UG/PG) and Working Professionals / Developers compete on equal terms across five problem statements.
-
Three-Phase Innovation Pipeline
-
An approximately 8-week Open Qualifying phase, an approximately 6-week PoC Development phase, and a 6–12 month Product Development phase for winning teams.
+
Two Challenge Phases
+
The Open Qualifying phase runs from 21 October to 21 December 2026. Shortlisted teams then advance to the PoC Development phase, which concludes on 5 March 2027.
One Winner per Problem Statement
-
Winners are selected after a live virtual demo and jury Q&A. Top performers advance to the Post-Challenge Product Development Grant.
+
One winner for each problem statement is selected after the final live virtual demo and jury Q&A.

Technology Theme: Applied AI on SoC

-

AI is moving from cloud-centric systems toward AI executed directly on System-on-Chips at the edge. This enables local processing, real-time response, privacy and operation without continuous connectivity, while meeting performance, power and cost constraints through the integration of AI software and semiconductor hardware.

+

Applied AI on SoC brings AI processing onto edge devices. Local processing can support real-time operation, privacy-aware use of data and offline capability, while teams consider performance, power and cost constraints.

The major shift in the 2nd Edition is from building AI applications to engineering AI-enabled products on SoCs.

-

The end-to-end engineering journey spans:

+

Participants will explore how AI models integrate with processors, memory, accelerators and software stacks. The end-to-end engineering journey spans:

-

Three-Phase Innovation Pipeline

+

Challenge Format

Phase 1 — Open Qualifying
-
Duration: ~8 weeks
-
Open qualifying stage where teams develop and submit their initial solution, report, code and demo.
+
21 October – 21 December 2026
+
Teams develop their initial solution and submit their report, source code and demo.
Phase 2 — PoC Development
-
Duration: ~6 weeks
-
Shortlisted teams receive mentorship and develop a fully deployed Proof of Concept.
-
-
-
Phase 3 — Product Development
-
Duration: 6–12 months
-
Winning teams receive continued technical, financial and ecosystem support to mature their PoCs into working prototypes / deployable products and accelerate go-to-market.
+
End of January – 5 March 2027
+
Shortlisted teams receive mentorship and develop and complete their proof of concept.
@@ -107,7 +106,7 @@

Project Problem Statements

Objective

Develop an AI-powered assistant that helps farmers identify crop diseases, detect pests and receive actionable advisory recommendations, running entirely on-device without internet connectivity.

Project Description

-

Students build an AI system capable of analysing crop images to:

+

Teams will build an AI system capable of analysing crop images to:

+ + + diff --git a/docs/Namma_Challenge_Page.html b/docs/Namma_Challenge_Page.html index 50235217..cd46ca91 100644 --- a/docs/Namma_Challenge_Page.html +++ b/docs/Namma_Challenge_Page.html @@ -23,15 +23,32 @@
-
Namma Smart AI Apps Challenge
-

Namma Smart AI Apps Challenge

-

AI-Native Smart Applications for Digital Karnataka

-

In collaboration with Arm, IIIT Bangalore & NASSCOM*

-
-
DurationAugust – November 2026
-
Team Size1–3 Members
-
ParticipantsStudents & Working Professionals / Developers
-
EligibilityIndian Nationals
+
+
+
Namma Smart AI Apps Challenge
+

Namma Smart AI Apps Challenge

+

AI-Native Smart Applications for Digital Karnataka & Digital India

+

In collaboration with Arm, IIIT Bangalore & NASSCOM

+
+
DurationSeptember – November 2026
+
Team Size1–3 Members
+
ParticipantsStudents & Working Professionals / Developers
+
EligibilityIndian Nationals
+
+
+
+ + + + + +
@@ -44,23 +61,22 @@

Namma Smart AI Apps Challenge

-

Program Overview

-

The Namma Smart AI Apps Challenge is a national-level initiative jointly launched by Arm and IIIT Bangalore with support of NASSCOM, to build a skilled talent pipeline in on-device Edge AI app development.

-

Unlike prior student-only skilling initiatives, the program is open to both student developers (UG/PG) and working professionals, recognising that Edge AI application-development talent spans academia and industry.

+

Programme Overview

+

The Namma Smart AI Apps Challenge is a national initiative from Arm and IIIT Bangalore, supported by NASSCOM, that helps participants build practical skills in Edge AI application development on edge devices.

+

The challenge is open to both students (UG/PG) and working professionals & developers, reflecting the breadth of Edge AI application-development talent across academia and industry.

+

The programme has a strong focus on skill development and provides project-based learning opportunities for all participants.

Participants will build real-world applications with societal impact across healthcare, education, telemedicine, gaming, and application modernisation, with mentorship from Arm engineers and IIIT Bangalore faculty.

-

The program officially launches on 15 August 2026 and culminates in a live demo day and winners' announcement at Bangalore Tech Summit 2026 on 19 November 2026.

-

Program Objectives

+

Programme Objectives

    -
  • Build practical, industry-relevant skills in on-device AI and app development.
    • -
  • Create a joint academia–industry talent pipeline by inviting both students and working professionals into the challenge.
    • -
  • Translate technology into real, deployable applications addressing genuine societal problems.
    • -
  • Establish IIIT Bangalore as an academic anchor for Edge AI curriculum, mentorship, and lab infrastructure.
    • -
  • Build a structured mentorship pipeline connecting participants directly with Arm engineers and IIIT Bangalore faculty.
    • -
  • Encourage cross-pollination between academic experimentation and industry engineering discipline.
    • -
  • Strengthen collaboration between Arm, NASSCOM, and IIIT Bangalore.
    • +
  • Build practical, industry-relevant skills in AI and application development on edge devices.
    • +
  • Develop real-world applications that address genuine societal challenges.
    • +
  • Gain hands-on experience in optimisation, benchmarking and deployment.
    • +
  • Learn directly from Arm engineers and IIIT Bangalore faculty through structured mentorship.
    • +
  • Gain experience working at the intersection of academic learning and industry engineering.
    • +
  • Strengthen practical understanding of Edge AI development on Arm-based platforms.
@@ -75,33 +91,35 @@

AI-Native Smart Applications for Digital Karnataka

What Makes This Challenge Different

This is an Edge AI challenge rather than a traditional cloud-only hackathon. Participants build AI-native on-device applications designed for privacy, low latency, responsiveness, offline capability, and power efficiency.

-

Real-world societal problem statements are paired with mentorship from Arm and academic guidance from IIIT Bangalore. Students and working professionals gain hands-on experience in building, optimisation, benchmarking, and deployment.

+

Real-world societal challenge areas are paired with mentorship from Arm and academic guidance from IIIT Bangalore. Students and working professionals gain hands-on experience in building, optimisation, benchmarking, and deployment.

-

Program Flow & Timeline

- - - - - - - - -
StageDateActivity
Launch15 August 2026Portal live and promotion via official handles
Registration15 August – 15 September 2026Team formation and selection of problem statements
Phase 1 — Ideation1 September – 30 September 2026Mentoring by experts, project baseline development, project report submission and demo video
Phase 1 Evaluation & Shortlist1 October – 10 October 2026Assessment of submitted projects by experts and shortlisting
Phase 2 — PoC Development11 October – 30 October 2026Full application development for shortlisted teams
Phase 2 Evaluation1 November – 15 November 2026Teams showcase a live virtual demo of their projects
Winners' Announcement19 November 2026Cash prize distribution at Bangalore Tech Summit 2026
+

Programme Flow & Timeline

+
+ + + + + + + + +
MilestoneTimingWhat Happens
Launch2nd September 2026The challenge portal opens
Registration2nd Sept – 20 Sept 2026Participants register their teams and select a challenge area
Phase 1 – Ideation20th – 30th Sept 2026Develop your initial solution with mentoring support and submit your report and demo video
Phase 1 Evaluation & Shortlist1 – 10 Oct 2026Submissions are evaluated and shortlisted teams are selected for Phase 2
Phase 2 – PoC Development11 – 30 Oct 2026Shortlisted teams develop their full proof-of-concept application
Phase 2 Evaluation1 – 15 Nov 2026Shortlisted teams present their solutions through a live virtual demo
Winners' Announcement19 Nov 2026Winners are announced at Bengaluru Tech Summit 2026
+

Phase 1 — Open Qualifying Round

-

All registered participants submit a project report with:

+

In Phase 1, all registered participants submit a project report containing:

-
Problem Statement Alignment
Clear articulation of the chosen problem and proposed solution approach.
+
Challenge Area Alignment
Clear articulation of the chosen challenge and proposed solution approach.
Technical Architecture
System design, model selection and Arm backend selection.
Working Code Repository
A public or shared GitHub repository with runnable code, README, setup instructions and at least a basic working prototype.
Performance Baseline
Initial benchmarking results or projections for on-device latency, memory footprint and accuracy.
Demo Video
A 3–5-minute video demonstrating the working prototype.
-
An expert panel evaluates Phase 1 submissions and shortlists the top teams per problem statement for Phase 2.
+
An expert panel evaluates Phase 1 submissions and shortlists the top teams per challenge area for Phase 2.
@@ -109,91 +127,228 @@

Phase 2 — App Development Round (Invitation Only)

Shortlisted teams receive hardware support, dedicated mentorship, and cloud compute credits to build a fully functional app.

Fully Deployed App
End-to-end on-device inference on target hardware, demonstrating all claimed use cases.
-
Optimization Report
Quantitative benchmarking across latency, memory and accuracy, with comparison of at least two backend/quantization configurations.
+
Optimisation Report
Quantitative benchmarking across latency, memory and accuracy, with comparison of at least two backend/quantisation configurations.
Updated Codebase
Production-quality code with documentation suitable for handover.
Live Virtual Demo
Presented to a jury panel with a Q&A session.
-

Project Problem Statements

-
1. AI Consultation Preparation Assistant for PatientsDomain: Healthcare
+

Project Challenge Areas

+
1. AI Consultation Preparation Assistant for Patients Domain: Healthcare

Domain: Healthcare
Applicable for: Students

-

In Karnataka's Tier-2 and Tier-3 towns and rural taluks, patients accessing online consultations under state public health initiatives such as Arogya Karnataka through Centre for Applied AI for Tech Solutions (CATS) often struggle to organise their symptoms, medical history and questions before the call, especially where the consultation is conducted in English while the patient is more comfortable in Kannada.

-

Build a desktop application that accepts a patient's symptoms, health notes, prescriptions and medical reports and generates:

-
  • a patient-friendly summary
  • relevant questions to ask the doctor
  • a consultation checklist
-

Kannada-language support must be treated as a core feature.

-

The application must run entirely on-device to protect sensitive health data and should be achievable by a student team using a small language model or ONNX-based text summarization pipeline.

+

In Karnataka's Tier-2 and Tier-3 towns and rural taluks, patients accessing online consultations can struggle to organise symptoms, medical history, prescriptions, reports and questions before a consultation, particularly when the consultation is in English and the patient is more comfortable in Kannada or any other Indian language.

+

Build an Arm64-native consultation-preparation application that uses AI to transform patient-provided notes and documents into a patient-friendly summary, relevant questions for the doctor and a consultation checklist, with Kannada or any other Indian language as a core experience.

+

Disclaimer: This solution is intended solely as a prototype/proof of concept for development, research and demonstration purposes. It is not medically validated, certified or approved and must not be used for diagnosis, treatment, clinical decision-making or real-world patient care.

+

The solution should prioritise privacy-conscious processing, useful offline/local workflows where feasible, and efficient Arm64 execution.

+

Students may develop, test and demonstrate the solution using the free Azure account and B2pts v2 Arm-based burstable VM allocation (up to 750 hours/month).

+

Physical WoA hardware, NPU access and SME2 are not prerequisites.

+

Teams should document which parts run locally on the Arm VM, any cloud services used, latency and resource usage, and how the architecture could map to a future Windows on Arm client.

+

Teams are encouraged to use ONNX Runtime for deploying and running AI models efficiently on Arm64 and explore Arm KleidiAI integration/optimisations, where applicable, to build an Arm-optimised AI software stack.

+

The Azure Arm64 VM can be used for development, model conversion, compilation, functional validation, inference and performance/resource analysis.

+

Teams should document:

+
    +
  • AI models and runtimes used
    • +
  • ONNX Runtime/KleidiAI integration where applicable
    • +
  • components running locally on the Arm VM
    • +
  • any external cloud services used
    • +
  • latency and resource utilisation
    • +
  • how the architecture could subsequently map to a Windows on Arm client
    • +
  • potential path to SME2-based optimisation on compatible hardware
    • +
+

Checklist

+
+ + +
Participant categoryAzure Arm64 accessAI inference optimisationCPU optimisationUse of KleidiAIPrimary technical validation
StudentsFree B2pts v2 Arm VM; $100 Azure credit for student account.RecommendedRecommendedOptional / where applicableArm64-native AI workflow; latency; CPU/memory efficiency; privacy/offline design; Windows readiness
+
+

Evaluation Criteria

+
+ + + + + + + +
Evaluation CriterionWeightWhat Judges Assess
Usefulness & Patient Experience20%Quality/relevance of summaries, questions and checklist; Kannada/other Indian language-first usability.
AI Output Quality & Latency20%Faithfulness, relevance and end-to-end response time on the stated Arm VM environment.
Privacy, Offline/Local Design15%Privacy-conscious architecture and degree to which core processing can operate locally or with limited connectivity.
Arm64-Native Implementation15%Arm64-native components, dependency choices and avoidance of unnecessary x64-only components.
Arm VM Efficiency & Reproducibility15%CPU/memory utilisation, responsiveness and repeatable measurements on the Azure Arm VM.
Engineering Quality & Windows Readiness15%Reliability, testing, packaging/design readiness and a credible path to a Windows on Arm client.
+
-
2. Personalized Study Guide, Planner and Knowledge Copilot for StudentsDomain: EdTech
+
2. Personalised Study Guide, Planner and Knowledge Copilot for Students Domain: EdTech

Domain: EdTech
Applicable for: Students

-

Students across Karnataka often have study material spread across notes, PDFs, lecture summaries and slides with no simple way to search it, test understanding or plan revision.

-

Build a desktop application that ingests a student's study material and allows the student to:

-
  • ask questions
  • generate flashcards
  • summarize topics
  • produce a personalized revision plan
-

The scope should remain lightweight and focus on a clean, usable local AI workflow rather than complex system integrations.

+

Students often work with study material scattered across notes, PDFs, lecture summaries and slides, with no simple way to search, test understanding or plan revision.

+

Build an Arm64-native study companion that ingests a student's material and uses AI to:

+
  • answer questions
  • summarise topics
  • generate flashcards or quizzes
  • create a personalised revision plan
+

The experience should remain useful with limited connectivity and should emphasise efficient Arm64 inference and a clean user workflow rather than complex integrations.

+

Students may use the free Azure account and B2pts v2 Arm-based burstable VM allocation (up to 750 hours/month) for development, testing and demonstration.

+

Physical WoA hardware, NPU access and SME2 are not mandatory.

+

Teams should report AI latency, throughput/resource utilisation and reproducible Arm64 measurements, and explain how the implementation can transition to a native Windows on Arm application.

+

Teams are encouraged to use ONNX Runtime for efficient AI model deployment and inference on Arm64 and explore Arm KleidiAI integration/optimisations, where applicable, to build an Arm-optimised AI software stack.

+

The Azure Arm64 VM can be used for model conversion, development, compilation, functional validation, inference, benchmarking and resource-utilisation analysis.

+

Teams should report:

+
    +
  • AI models and runtimes used
    • +
  • ONNX Runtime/KleidiAI integration where applicable
    • +
  • inference latency
    • +
  • throughput
    • +
  • resource utilisation
    • +
  • reproducible Arm64 measurements
    • +
  • path to a native Windows on Arm client
    • +
  • potential path to SME2-based optimisation on compatible hardware
    • +
+

Checklist

+
+ + +
Participant categoryAzure Arm64 accessAI inference optimisationCPU optimisationUse of KleidiAIPrimary technical validation
StudentsFree B2pts v2 Arm VM; $100 Azure credit for student account.RecommendedRecommendedOptional / where applicableArm64-native AI workflow; latency/throughput; CPU/memory efficiency; local study-data workflow
+
+

Evaluation Criteria

+
+ + + + + + + +
Evaluation CriterionWeightWhat Judges Assess
Learning Usefulness & UX20%Effectiveness of Q&A, summaries, flashcards/quizzes and revision planning.
AI Quality & Latency20%Grounding in supplied study material and measured response time.
Offline / Limited-Connectivity Capability15%Usefulness when network access is limited and sensible data handling.
Arm64-Native Implementation15%Native Arm64 build/dependencies and efficient execution.
Arm VM Performance15%Throughput, CPU/memory utilisation and reproducible performance evidence.
Engineering Quality & Windows Readiness15%File ingestion, reliability, testing, documentation and path to native Windows on Arm.
+
-
3. Telemedicine Consultation Copilot with Local AI Summarization and Clinical Workflow SupportDomain: Telemedicine
-

Domain: Telemedicine
Applicable for: Working Professionals / Developers

-

Karnataka's public health system includes online consultations connecting patients in rural districts to doctors at taluk and district hospitals.

-

Clinicians spend significant time documenting symptoms, medications, investigations and follow-up actions.

-

Develop a desktop application that:

-
  • captures or imports consultation transcripts
  • extracts relevant clinical details
  • generates a structured clinician-facing note
  • generates a separate patient-friendly summary
-

Kannada must be supported as a first-class language.

-

Advanced teams may extend the solution with:

-
  • additional regional-language support
  • retrieval from medical reference documents
  • structured follow-up planning
+
3. Telemedicine Consultation Copilot with Local AI Summarisation and Clinical Workflow Support Domain: Healthcare
+

Domain: Healthcare
Applicable for: Working Professionals / Developers

+

Clinicians conducting remote consultations can spend significant time documenting symptoms, medications, investigations and follow-up actions, reducing time available for patient care and increasing the risk of omissions.

+

Develop an Arm64-native telemedicine consultation copilot that captures or imports consultation transcripts, extracts relevant clinical details and generates both:

+
  • a structured clinician-facing note
  • a separate patient-friendly summary
+

Disclaimer: This solution is intended solely as a prototype/proof of concept for development, research and demonstration purposes. It is not medically validated, certified or approved and must not be used for diagnosis, treatment, clinical decision-making or real-world patient care.

+

Kannada/other Indian language should be a first-class language.

+

Optional extensions include:

+
  • additional regional languages
  • retrieval from medical reference documents
  • structured follow-up planning
+

Because a physical WoA device is not assumed, Phase 1 should be developed and validated on the free Azure B2pts v2 Arm-based VM environment (up to 750 hours/month).

+

Shortlisted Phase 2 teams may use the stated $200 Azure credit to scale or deepen the PoC.

+

SME2 is not required or scored.

+

Evaluation should focus on available Arm64-native execution and measurable CPU/resource performance rather than unavailable device-specific NPU/battery metrics, including:

+
  • Arm64-native execution
  • output quality
  • latency
  • reproducibility
  • privacy-conscious architecture
  • CPU/resource optimisation
  • practical workflow usefulness
+

Teams are encouraged to use ONNX Runtime for efficient AI model deployment and inference on Arm64 and explore Arm KleidiAI integration/optimisations, where applicable, to build an Arm-optimised AI software stack.

+

The Azure Arm64 VM may be used for model conversion, development, compilation, functional validation, inference, benchmarking and resource-utilisation analysis.

+

Teams should also document:

+
  • AI models and runtimes
  • ONNX Runtime/KleidiAI integration where applicable
  • transition path to a Windows on Arm client
  • potential path to SME2-based optimisation on compatible hardware
+

Checklist

+
+ + +
Participant categoryAzure Arm64 accessAI inference optimisationCPU optimisationUse of KleidiAIPrimary technical validation
Working Professionals / DevelopersFree B2pts v2 Arm VM; $200 Azure credit for new account.RecommendedRequiredOptional / where applicableArm64-native execution; CPU optimisation; AI latency/throughput/memory; reproducibility; clinical workflow quality
+
+

Evaluation Criteria

+
+ + + + + + + +
Evaluation CriterionWeightWhat Judges Assess
Clinical Workflow Usefulness & Output Quality25%Completeness and practical usefulness of extracted details, clinician note and patient summary; no diagnosis requirement.
AI Accuracy & Latency20%Faithfulness to consultation input, extraction/summarisation quality and measured latency.
Language, Privacy & Local Design15%Kannada/other Indian language-first workflow, optional regional languages and privacy-conscious processing.
Arm64-Native Implementation & CPU Optimisation15%Native Arm64 components, optimised CPU execution and dependency quality.
Azure Arm Performance & Reproducibility10%Repeatable latency/throughput/memory/CPU measurements on the provided Arm VM.
Engineering Quality & Scalability15%Robustness, testing, documentation, deployment quality and effective use of Phase 2 Azure resources.
+
-
4. Real-Time AI Gaming Coach and Multimedia Highlights StudioDomain: Gaming
+
4. Real-Time AI Gaming Coach and Multimedia Highlights Studio Domain: Gaming

Domain: Gaming
Applicable for: Working Professionals / Developers

-

Build a desktop application that analyses recorded gameplay clips or live screen/video input and:

-
  • detects key events
  • tags important moments
  • generates coaching insights
  • generates highlight summaries
  • combines computer vision with audio/text metadata
  • presents findings through a local assistant interface
  • surfaces mistakes, achievements and strategy recommendations
-

This problem targets experienced developers because it requires real-time video processing, multi-model orchestration and a performance-aware desktop UX.

+

Build an Arm64-native AI gaming and multimedia analysis application that processes recorded gameplay clips or suitable streamed/video inputs, detects and tags key events, and generates coaching insights and highlight summaries through an assistant-style experience.

+

The pipeline may combine computer vision with audio/text metadata and should be profiled and optimised for Arm64 execution.

+

Phase 1 should run on the free Azure B2pts v2 Arm-based VM environment (up to 750 hours/month), with the stated $200 Azure credit available to shortlisted teams in Phase 2.

+

SME2, NPU utilisation, battery measurements and device-only real-time capture are not mandatory.

+

Teams should instead demonstrate measurable improvements through reproducible Arm64 profiling, for example:

+
  • inference latency
  • throughput
  • CPU utilisation
  • memory footprint
  • processing time per minute of video
  • cost-efficiency
+

Use a clearly defined baseline where a valid comparison can be produced.

+

Teams are encouraged to use ONNX Runtime for efficient AI model deployment and inference on Arm64 and explore Arm KleidiAI integration/optimisations, where applicable, to build an Arm-optimised AI software stack.

+

The Azure Arm64 VM may be used for model conversion, development, compilation, functional validation, inference, benchmarking and resource-utilisation analysis.

+

Teams should also document:

+
  • AI models and runtimes
  • ONNX Runtime/KleidiAI integration where applicable
  • path to a Windows on Arm client
  • potential path to SME2-based optimisation on compatible hardware
+

Checklist

+
+ + +
Participant categoryAzure Arm64 accessAI inference optimisationCPU optimisationUse of KleidiAIPrimary technical validation
Working Professionals / DevelopersFree B2pts v2 Arm VM; $200 Azure credit for new account.RecommendedRequiredOptional / where applicableArm64 profiling; media/AI throughput; latency; CPU/memory efficiency; repeatable baseline comparison where valid
+
+

Evaluation Criteria

+
+ + + + + + + +
Evaluation CriterionWeightWhat Judges Assess
Arm64 Performance Improvement25%Measured improvement against a clearly defined and valid baseline using cloud-accessible metrics.
AI / Multimedia Pipeline Quality20%Robust event detection, tagging, coaching insights and highlight generation.
Throughput & Resource Efficiency15%Inference latency, throughput, CPU utilisation, memory footprint and processing efficiency.
User Experience & Usefulness15%Actionable output and clear value to gamers/creators; demo usability.
Native Arm64 Completeness10%Arm64-native application/pipeline and dependencies where feasible.
Engineering & Benchmark Reproducibility15%Repeatable methodology, profiling evidence, testing, reliability and documentation.
+
-
5. Desktop Application Migration & Optimization with Performance Benchmarking SuiteDomain: App Migration
-

Domain: App Migration
Applicable for: Students & Working Professionals / Developers

-

Select an existing open-source or self-built x86 Windows desktop application and migrate it to run natively on Windows on Arm (Arm64).

-

Address real migration issues where relevant, including:

-
  • unavailable or x86-only dependencies
  • native library recompilation
  • driver or SDK incompatibilities
  • installer or packaging changes
-

The deliverable must include both:

  1. a working native Arm64 port
  2. a rigorous and repeatable benchmarking suite
-

The benchmarking suite must compare x86-emulated and native Arm64 versions across:

  • performance
  • power consumption
  • memory utilisation
-

Student teams may choose a smaller, self-contained application such as a utility, lightweight productivity app or sample ONNX/AI inference app.

-

Professional teams are encouraged to attempt a larger, dependency-heavy or AI-inference-capable application demonstrating production complexity.

+
5. Desktop Application Migration & Optimisation with Performance Benchmarking Suite Domain: App Migration / Open Source / Performance
+

Domain: App Migration / Open Source / Performance
Applicable for: Students & Working Professionals / Developers

+

Select an existing, meaningful Windows application, open-source library, runtime, framework, SDK or developer tool that people or developers actually use and advance it toward a complete native Arm64 experience.

+

Teams may pursue:

+

Port It Native

+

Port a popular application or component to Arm64.

+

Open Source Hero / Unlock the Ecosystem

+

Make an important dependency or developer technology Arm64-native and unlock downstream applications.

+

Faster on Arm

+

Optimise an existing workload and demonstrate a meaningful native Arm64 performance improvement.

+

Resolve real migration issues such as:

+
  • x86/x64-only dependencies
  • native library recompilation
  • build-system changes
  • CI/testing
  • SDK compatibility
  • installers
  • packaging
+

Students may use the free Azure account with B2pts v2 Arm-based burstable VM access (up to 750 hours/month).

+

Working professionals/developers use the same free Arm VM environment in Phase 1 and may use the stated $200 Azure credit in Phase 2.

+

Where Windows-specific GUI or installer behaviour cannot be exercised on the provided Azure Arm VM, teams should clearly separate:

+
  • what has been implemented
  • what has been cross-compiled
  • what has been CI-validated
  • what remains device-dependent
+

Judging should reward:

+
  • native-port completeness
  • native dependencies
  • ecosystem impact
  • technical difficulty
  • implementation quality
  • reproducible Arm64 benchmarking
  • CI/testing
  • upstream contribution
  • user/developer impact
+

For AI/ML applications, runtimes or frameworks, teams are encouraged to use ONNX Runtime for efficient Arm64 inference and explore Arm KleidiAI integration/optimisations where applicable.

+

ONNX Runtime/KleidiAI usage should be applied only where relevant and is not mandatory for non-AI software ports.

+

For applicable AI workloads, teams should additionally demonstrate and document:

+
  • ONNX Runtime/KleidiAI integration
  • inference performance
  • CPU/resource optimisation
+

Checklist

+
+ + + +
Participant categoryAzure Arm64 accessAI inference optimisationCPU optimisationUse of KleidiAIPrimary technical validation
StudentsFree B2pts v2 Arm VM; $100 Azure credit for new account.As applicable to selected projectRequiredOptional / if AI workloadNative Arm64 port/dependencies; build & CI; reproducible benchmarks; ecosystem impact
Working Professionals / DevelopersFree B2pts v2 Arm VM; $200 Azure credit for new account.As applicable to selected projectRequiredOptional / if AI workloadNative Arm64 port/dependencies; CI/testing; reproducible benchmarks; upstream acceptance & ecosystem impact
+
+

Evaluation Criteria

+
+ + + + + + + +
Evaluation CriterionWeightWhat Judges Assess
Completeness of Native Arm64 Port / Enablement20%Functional completeness and reduction/elimination of x86/x64-only dependencies within the accessible environment.
Ecosystem / User Impact20%Importance/popularity of project and downstream users/apps/developers unlocked.
Arm64 Performance Evidence15%Repeatable startup, throughput, latency, memory, CPU or other relevant metrics; valid baseline required for comparison claims.
Technical Difficulty & Native Dependencies15%Depth of dependency, build-system, SDK, CI, packaging or migration work.
Implementation Quality, CI & Testing15%Maintainability, correctness, automated Arm64 testing, documentation and reproducibility.
Upstream Acceptance & Sustainability15%Quality of upstream contribution: accepted/merged changes receive strongest credit; submitted Pull Requests/issues with maintainer engagement and a credible upstream path also receive credit.
+
+
+

Azure VM availability is subject to Microsoft subscription, region, quota and capacity restrictions. Teams should use the Azure Free Services configuration where available and report any subscription or regional limitations to the organisers.

+

Teams are not expected to run large foundation models entirely on the B2pts v2 VM. Lightweight, quantised, distilled or otherwise resource-efficient models are encouraged. Cloud AI services may be used where appropriate, but Arm64-local components must be clearly identified and benchmarked where required.

+

Eligibility Criteria

The challenge is open to Indian nationals in two categories.

- +
CategoryWho qualifiesTeam Lead Requirement
StudentsUG or PG students enrolled in a regular full-time course at a recognized Indian HEI with a background in AI/ML, digital systems, or software.Team Leader must be an enrolled student. Faculty Mentor required.
StudentsUG or PG students enrolled in a regular full-time course at a recognised Indian HEI with a background in AI/ML, digital systems, or software.Team Leader must be an enrolled student. Faculty Mentor required.
Working Professionals / DevelopersIndividuals currently employed in industry, startups, or independent developers with demonstrable technical experience in AI/ML or embedded systems.Team Leader is a professional. Faculty Mentor not required.

Team Rules

    -
  • Team size: 1 to 3 members per team.
  • Students: All members must be from the same institution and campus; they may span different branches and semesters.
  • Faculty Mentor: Each student team should have a Faculty Mentor.
  • Professional teams: Members may be from different organizations.
  • Problem statement: Each team selects one problem statement.
  • One change of team composition or problem statement is permitted within 15 days of registration.
    • +
  • Team size: 1 to 3 members per team.
  • Students: All members must be from the same institution and campus; they may span different branches and semesters.
  • Faculty Mentor: Each student team should have a Faculty Mentor.
  • Professional teams: Members may be from different organisations.
  • Challenge area: Each team selects one challenge area.
  • One change of team composition or challenge area is permitted within 15 days of registration.
-
-

Problem Statement Compliance Checklist

-
- - - - - - -
PS #DomainParticipant CategoryFree Azure AccountWoA DeviceSVE2SME2Inference AccelerationCPU OptimizationKleidiAI
1HealthcareStudentsMandatoryOptional / RecommendedMandatoryOptional / RecommendedOptional / RecommendedOptional / RecommendedOptional / Recommended
2EdTechStudentsMandatoryOptional / RecommendedMandatoryOptional / RecommendedOptional / RecommendedOptional / RecommendedOptional / Recommended
3TelemedicineWorking Professionals / DevelopersNot RequiredMandatoryNot RequiredMandatoryMandatoryMandatoryMandatory
4GamingWorking Professionals / DevelopersNot RequiredMandatoryNot RequiredMandatoryMandatoryMandatoryMandatory
5App MigrationStudentsMandatoryOptional / RecommendedMandatoryOptional / RecommendedOptional / RecommendedOptional / RecommendedOptional / Recommended
5App MigrationWorking Professionals / DevelopersNot RequiredMandatoryNot RequiredMandatoryMandatoryMandatoryMandatory
-
Challenge 5 is open to both Students and Working Professionals / Developers. Each participant category is held to its respective track's compliance requirements.
-
-

Recognition & Rewards

    -
  • Participation Certificates for all participants submitting a working project.
  • Arm-branded merchandise for the Top 40 finalists.
  • Cash awards for winning teams across each of the five problem statements.
  • A few of the top winning projects may be featured at national-level industry conferences.
    • +
  • Participants who submit a working project will receive a certificate of participation.
  • Arm-branded merchandise for all finalists.
  • IIT Delhi will provide a cash award of ₹1 lakh to each winning team, with one winning team selected for each challenge area.
  • The Top 20 finalists will have the opportunity to enrol in advanced Edge AI programmes involving IIIT Bangalore, NASSCOM CoE and Arm, or to have their projects from this challenge showcased at national-level industry conferences.

Registration

- Complete the registration form below to register for the + Complete the form below to register for the Namma Smart AI Apps Challenge.

@@ -213,14 +368,20 @@

Registration

2. How many people can be in a team?

1 to 3 members.

3. Can students from different institutions form a team?

No. Student team members must be from the same institution and campus, although they may come from different branches and semesters.

4. Do student teams need a Faculty Mentor?

Yes. Each student team should have a Faculty Mentor.

-
5. Can professionals from different organizations form a team?

Yes.

-
6. Which challenges can students enter?

Challenges 1, 2 and 5.

-
7. Which challenges can professionals/developers enter?

Challenges 3, 4 and 5.

-
8. Can a team change its composition or problem statement?

One change of team composition or problem statement is permitted within 15 days of registration.

-
9. What happens in Phase 1?

Participants submit problem statement alignment, technical architecture, a working code repository, a performance baseline and a 3–5-minute demo video.

+
5. Can professionals from different organisations form a team?

Yes.

+
6. Which challenge areas can students enter?

Challenge Areas 1, 2 and 5.

+
7. Which challenge areas can professionals/developers enter?

Challenge Areas 3, 4 and 5.

+
8. Can a team change its composition or challenge area?

One change of team composition or challenge area is permitted within 15 days of registration.

+
9. What happens in Phase 1?

Participants submit challenge area alignment, technical architecture, a working code repository, a performance baseline and a 3–5-minute demo video.

10. What support do shortlisted teams receive?

Hardware support, dedicated mentorship and cloud compute credits.

-
11. When are winners announced?

19 November 2026 at Bangalore Tech Summit 2026.

-
12. What rewards are available?

Participation certificates for participants submitting a working project, Arm-branded merchandise for the Top 40 finalists, cash awards for winning teams across all five problem statements, and potential national-level conference exposure for some top winning projects. Exact cash award amounts are not currently specified.

+
11. When are winners announced?

19 November 2026 at Bengaluru Tech Summit 2026.

+
12. What rewards are available?

Participants who submit a working project will receive a certificate of participation. All finalists will receive Arm-branded merchandise. IIT Delhi will provide a cash award of ₹1 lakh to each winning team, with one winning team selected for each challenge area. The Top 20 finalists will also have the opportunity to enrol in advanced Edge AI programmes involving IIIT Bangalore, NASSCOM CoE and Arm, or to have their projects from this challenge showcased at national-level industry conferences.

+
13. How should participants handle materials and intellectual property related to their project submission?
+

Participants are responsible for ensuring that they have all necessary rights, permissions and authorisations to use and submit any intellectual property, software, code, data, materials or other content used in connection with their participation in the Namma Smart AI Apps Challenge (“Participant Materials”).

+

Participants must not submit or disclose any confidential, proprietary or third-party information unless they are authorised to do so and such disclosure does not breach any confidentiality, employment, contractual or other obligation owed to any third party.

+

Participation in the Challenge does not require participants to disclose their own confidential information. Any Participant Materials voluntarily submitted or disclosed as part of the Challenge are provided at the participant’s own discretion and responsibility.

+

Each participant remains responsible for ensuring that their participation and submission do not infringe, misappropriate or otherwise violate any third-party intellectual property or other rights.

+
diff --git a/docs/_sass/components/_challengePage.scss b/docs/_sass/components/_challengePage.scss index c62d5904..837ac8d3 100644 --- a/docs/_sass/components/_challengePage.scss +++ b/docs/_sass/components/_challengePage.scss @@ -309,7 +309,7 @@ html { .ch-accordion summary::after { content: "▾"; margin-left: auto; - font-size: 12px; + font-size: 30px; opacity: 0.6; transition: transform 0.2s ease; } @@ -570,6 +570,76 @@ html { z-index: 1; } +.namma-challenge .ch-hero-layout { + display: grid; + grid-template-columns: minmax(0, 1fr) minmax(480px, 0.9fr); + align-items: center; + gap: 24px; +} + +.namma-challenge .ch-hero-copy { + min-width: 0; +} + +.namma-challenge .ch-logo-row { + display: flex; + flex-flow: row wrap; + align-items: center; + justify-content: flex-end; + gap: 6px; + min-width: 0; +} + +.namma-challenge .ch-logo { + display: inline-flex; + align-items: center; + justify-content: center; + flex: 0 0 auto; + + height: auto; + padding: 0; + + border: 0; + background: transparent; + box-shadow: none; + + overflow: visible; +} + +.namma-challenge .ch-logo img { + display: block; + width: auto; + height: auto; + + max-height: 74px; + + object-fit: contain; + border-radius: 10px; +} + +.namma-challenge .ch-logo--wide img { + max-width: 330px; +} + + + +.namma-challenge .ch-logo--iiitb img { + max-height: 150px; +} + +@media (max-width: 1040px) { + .namma-challenge .ch-hero-layout { + grid-template-columns: minmax(0, 1fr) minmax(340px, 0.85fr); + gap: 18px; + } + + .namma-challenge .ch-logo-row { + justify-content: center; + max-width: 360px; + margin-left: auto; + } +} + .namma-challenge .ch-badge-dot { background: var(--purple-accent); } @@ -695,7 +765,65 @@ html { font-weight: 800; } +.namma-challenge .ch-problem-statements .ch-acc-body > h3 { + margin: 20px 0 8px; + color: var(--purple-dark); + font-size: 16px; +} + +.namma-challenge .ch-table-scroll { + max-width: 100%; + overflow-x: auto; + border: 1px solid var(--purple-border); + border-radius: 8px; + -webkit-overflow-scrolling: touch; +} + +.namma-challenge .ch-table-scroll:focus-visible { + outline: 3px solid rgba(3, 102, 214, 0.28); + outline-offset: 2px; +} + +.namma-challenge .ch-table-scroll .ch-table { + margin-top: 0; +} + +.namma-challenge .ch-ps-table { + min-width: 760px; +} + +.namma-challenge .ch-ps-checklist { + min-width: 980px; +} + +.namma-challenge .ch-ps-table th, +.namma-challenge .ch-ps-table td { + line-height: 1.4; + vertical-align: top; +} + +.namma-challenge .ch-ps-evaluation th:nth-child(1), +.namma-challenge .ch-ps-evaluation td:nth-child(1) { + width: 30%; +} + +.namma-challenge .ch-ps-evaluation th:nth-child(2), +.namma-challenge .ch-ps-evaluation td:nth-child(2) { + width: 10%; + text-align: center; + white-space: nowrap; +} + +.namma-challenge .ch-ps-evaluation th:nth-child(3), +.namma-challenge .ch-ps-evaluation td:nth-child(3) { + width: 60%; +} + /* Timeline emphasis without changing the table structure. */ +.namma-challenge .ch-timeline { + min-width: 700px; +} + .namma-challenge .ch-timeline td:first-child { color: var(--purple-dark); font-weight: 700; @@ -823,6 +951,30 @@ html { margin: 8px 0; } + .namma-challenge .ch-hero-layout { + grid-template-columns: minmax(0, 1fr); + gap: 20px; + } + + .namma-challenge .ch-hero-layout > .ch-logo-row { + flex-direction: row !important; + justify-content: center; + max-width: none; + margin-left: 0; + } + + .namma-challenge .ch-logo { + height: 68px; + } + + .namma-challenge .ch-hero-layout .ch-logo-row .ch-logo img { + width: auto !important; + height: auto !important; + max-height: 52px !important; + padding: 0 !important; + border-radius: 0 !important; + } + .ch-tagline { font-size: 16px; line-height: 1.5; @@ -1010,6 +1162,22 @@ html { font-size: 20px; } + .namma-challenge .ch-hero-layout > .ch-logo-row { + flex-direction: row !important; + } + + .namma-challenge .ch-logo { + max-width: 100%; + } + + .namma-challenge .ch-hero-layout .ch-logo-row .ch-logo img { + max-height: 48px !important; + } + + .namma-challenge .ch-logo--wide img { + max-width: min(252px, calc(100vw - 74px)); + } + .ch-tagline { font-size: 15px; line-height: 1.5; @@ -1250,12 +1418,17 @@ html { color: var(--text-muted); } -.bharat-2026-challenge .bharat-whats-new .phase-deliverables, -.bharat-2026-challenge .bharat-pipeline .phase-deliverables { +.bharat-2026-challenge .bharat-whats-new .phase-deliverables { grid-template-columns: repeat(3, minmax(0, 1fr)); align-items: stretch; } +.bharat-2026-challenge .bharat-pipeline .phase-deliverables { + grid-template-columns: repeat(2, minmax(0, 316px)); + justify-content: center; + align-items: stretch; +} + .bharat-2026-challenge .bharat-whats-new .phase-deliverable { border-top: 3px solid rgba(182, 101, 38, 0.65); } @@ -1370,6 +1543,23 @@ html { border-radius: 9px; } +.bharat-2026-challenge .bharat-faq .ch-accordion summary { + display: flex; + flex-direction: row; + min-height: 0; + padding: 14px 48px 14px 16px; +} + +.bharat-2026-challenge .bharat-faq .ch-acc-title { + display: block; + min-width: 0; + color: var(--text-dark); + font-size: 16px; + line-height: 1.4; + letter-spacing: normal; + text-transform: none; +} + .bharat-2026-challenge .ch-acc-title { display: contents; } @@ -1543,6 +1733,67 @@ html { box-shadow: inset 0 0 0 1px rgba(239, 191, 117, 0.18); } +.bharat-2026-challenge .ch-register-cta { + margin: -44px 0 24px; +} + +.bharat-2026-challenge #registration { + scroll-margin-top: 100px; +} + +.bharat-2026-challenge .bharat-registration { + border-color: rgba(217, 154, 61, 0.58); + background: + radial-gradient(circle at 92% 12%, rgba(217, 154, 61, 0.14), transparent 30%), + linear-gradient(135deg, #20211e, #2c2b26); + box-shadow: inset 4px 0 0 #d99a3d, + 0 10px 24px rgba(32, 32, 29, 0.18); +} + +.bharat-2026-challenge .bharat-registration > h2 { + color: #fffaf2; +} + +.bharat-2026-challenge .bharat-registration > h3 { + color: #efbf75; +} + +.bharat-2026-challenge .bharat-registration > p, +.bharat-2026-challenge .bharat-registration > ul li { + color: #f5eee4; +} + +.bharat-2026-challenge .registration-form-wrapper { + border-color: rgba(217, 154, 61, 0.55); + background: #fffdfa; + box-shadow: inset 4px 0 0 rgba(217, 154, 61, 0.72), + 0 8px 20px rgba(0, 0, 0, 0.16); +} + +.bharat-2026-challenge .registration-form-wrapper input[type="text"], +.bharat-2026-challenge .registration-form-wrapper input[type="email"], +.bharat-2026-challenge .registration-form-wrapper input[type="tel"], +.bharat-2026-challenge .registration-form-wrapper select, +.bharat-2026-challenge .registration-form-wrapper textarea { + border-color: #c9ad91 !important; +} + +.bharat-2026-challenge .registration-form-wrapper input:focus, +.bharat-2026-challenge .registration-form-wrapper select:focus, +.bharat-2026-challenge .registration-form-wrapper textarea:focus { + outline: 2px solid rgba(182, 101, 38, 0.3) !important; + border-color: #b66526 !important; +} + +.bharat-2026-challenge .registration-form-wrapper .mktoButton:hover { + background: #b66526 !important; +} + +.bharat-2026-challenge .registration-form-wrapper .mktoButton:focus-visible { + outline: 3px solid rgba(217, 154, 61, 0.46); + outline-offset: 3px; +} + .bharat-2026-challenge .ch-register-button { border: 2px solid #292823; background: #d99a3d; diff --git a/docs/images/Arm_Brand_Logo_Guidelines_300x175px.jpg b/docs/images/Arm_Brand_Logo_Guidelines_300x175px.jpg new file mode 100644 index 00000000..a855c85f Binary files /dev/null and b/docs/images/Arm_Brand_Logo_Guidelines_300x175px.jpg differ diff --git a/docs/images/IIITB_logo1.jpg b/docs/images/IIITB_logo1.jpg new file mode 100644 index 00000000..f3b89b18 Binary files /dev/null and b/docs/images/IIITB_logo1.jpg differ diff --git a/docs/images/Logo-CATS-GOK.jpg b/docs/images/Logo-CATS-GOK.jpg new file mode 100644 index 00000000..ab4c7ea8 Binary files /dev/null and b/docs/images/Logo-CATS-GOK.jpg differ