ExpertFingerprinting: Behavioral Pattern Analysis and Specialization Mapping of Experts in GPT-OSS-20B's Mixture-of-Experts Architecture

A Java client to interact with Arize API

Model audit software

Capability Schema Spec defines a shared semantic language for world model evaluation. Standardize capability definition, observation, and verification across models and benchmarks. Not a benchmark—a shared language. Define • Observe • Verify

Architecture and training decisions determine how observable an LLM is. Transformer activations carry decision-quality signals that output confidence misses; training can preserve or erase them during convergence, even as predictive performance improves.

Reference implementation of the Capability Schema Specification. Proves that world model capabilities can be defined, observed, and verified in practice — with real checkpoints, real simulators, and real scores. Define • Observe • Verify • Deliver

ML Monitoring System – Production-Ready Model Observability Platform A Flask-based ML observability system designed to track model performance, detect drift, and monitor real-time prediction behavior.

How to observe LLMs using Traceloop

"An end-to-end Medical Imaging pipeline built on AWS SageMaker utilizing Transfer Learning (ResNet18). The project implements Hyperparameter Optimization (HPO) to minimize loss, leverages SageMaker Debugger & Profiler for resource optimization, and concludes with a Production-ready real-time inference endpoint

Statistical drift detection service for production ML. PSI, KS and Jensen-Shannon detectors behind a FastAPI scoring API, with Docker, CI and reproducible benchmarks.

Token-level expert routing capture for Nemotron-Cascade-2-30B-A3B MoE layers during vLLM inference. Parquet output.

Production-style model reliability platform with drift checks, durable incidents, transactional notifications, RCA and lineage, a tested SignalOps console, and a study guide.