Applied AI & Open-Source Labs
Applied proof-of-concepts and open-source tooling. Combining production data engineering discipline with academic Information Retrieval (UIUC CS 410 / CS 441) and Agentic AI workflows (Model Context Protocol).
Building an MCP Server for Data Pipeline Triage & BigQuery Optimization
Empowering AI coding assistants with Model Context Protocol (MCP) tools to autonomously inspect BigQuery query execution plans, parse Airflow failure logs, and suggest partitioning and clustering fixes.
Hybrid Search & BM25 + Dense Retrieval for Enterprise Data Catalogs
Why naive vector search fails on technical schemas and error logs, and how combining Okapi BM25 lexical ranking with dense semantic vectors (RRF) builds bulletproof RAG for Data Engineers.
Automated YouTube & Media Asset Pipeline: FastAPI, DuckDB, Polars & Local LLMs
A production-grade local automation pipeline transforming raw video transcripts, timestamps, and slide benchmarks into SEO-optimized YouTube chapters, metadata, and cross-platform markdown notes.
Upcoming YouTube Breakdowns & Repos
Each lab is accompanied by a modular GitHub repository and a deep-dive YouTube video walkthrough explaining the code and architectural trade-offs.