At InfoBay.AI, I build Python pipelines for AI datasets, including classification, infographic detection, AI-generated content detection, image enhancement, and OCR. I design inference workflows with routing, fallback paths, and audit logging.
I developed an infographic-detection approach combining visual entropy, visual density, word spacing, and selective OCR. I also reduced a manual image-redaction task from about five days to one using fixed-coordinate repainting and an OCR plus regex PII detector.
I built an AI Knowledge Platform with hybrid BM25 and semantic retrieval, reranking, conversational memory, and LangGraph workflows. MCP servers connect its agents to tools and external knowledge.
My other projects include Reliability-Lab, an AI evaluation and regression-testing framework, and Zwip-AI, a full-stack platform with AI tools. I’ve also contributed fixes to Jenkins MCP Server Plugin and Unchained-Caravan.

