$ whoami
> Thanish Saravanakumar — Applied AI & Cloud Engineer

$ cat focus.txt
> LLMs · RAG pipelines · Computer Vision · Cloud-native ML systems

$ ./currently.sh
> Shipping production LLM apps for 5,000+ daily users, 8+ languages   # @ Infosys BPM

I'm drawn to AI that can show its work — a RAG answer with sources, a diagnosis with an explanation, a detection you can actually trust.

I build production AI systems end to end — RAG pipelines, LLM applications, and the FastAPI microservices that serve them at scale. Before that, I spent four months at DRDO's Centre for Airborne Systems building real-time object detection for aerial surveillance, which is where the research habit started. Currently an M.Tech candidate in AI & Data Science, still treating every production system like an open research question.

_{Plain chips are practices. Two-tone chips pin the actual library version I run.}

AI / ML

Backend, Infra & Cloud

Data

Languages & Tools

[1] ResearcherAI — live demo ↗ Agentic research platform with federated search and citation-grounded RAG chat for literature review. RAG Agents LLM

[2] Ovarian Cancer Diagnosis — source ↗ Explainable ML clinical decision-support system for diagnosis assistance, built with SHAP-based interpretability. ML Healthcare XAI

AI Engineer — Infosys BPM Ltd. _{Jul 2025 – Present}

• Building LLM applications serving 5,000+ daily users
• RAG pipelines spanning 8+ languages
• FastAPI microservices handling 50K+ requests/day

AI & ML Intern — DRDO, Centre for Airborne Systems _{Jan – Apr 2025}

• Real-time object detection for aerial surveillance, 92%+ accuracy
• Optimized edge deployment, cutting latency by 35%

Advanced Data Analytics Intern — GDS, AICTE _{Feb – Apr 2024}

• Analyzed healthcare datasets and built Power BI dashboards

Currently reading up on agentic RAG evaluation and edge deployment for vision models — always up for a conversation about either.

— Thanish Saravanakumar