Built in the Open

RAG evaluation toolkit that generates synthetic Q&A datasets you can actually trust — a 6-stage filter pipeline removes hallucinated, ambiguous, and trivial pairs before they reach your benchmark.

Reference implementation of a permission-aware hybrid retrieval pipeline — dense + sparse fusion, reranking, and citation-grounded generation.

An interactive, always-current roadmap for learning GenAI — from foundations to production systems, with curated resources at every step.

High-throughput NLP pipeline classifying human sentiment during COVID-19 using classical ML — Logistic Regression, Naive Bayes, Random Forest, and Decision Trees.

Automated insurance claim lifecycle with real-time forecasting of approval timelines using Naive Bayes over volatile datasets.

Secure full-stack engagement ecosystem with multi-channel registration, rate limiting, and a 30+ module admin command center — live at GCETTB.