Vedant Kugaonkar

I’m an AI/ML Engineer who builds end-to-end systems that work in real environments—especially where computer vision and LLMs meet production-grade performance. In my AI and Computer Vision Internship at Engineering Bay Corporation, I designed and deployed a production-grade YOLOv8-based UAV vision system in Python, reaching 95% real-time accuracy for multi-class object detection and improving autonomous obstacle detection results by 40% in field tests.

I also focus on modern language capabilities through Retrieval-Augmented Generation. I architected LawBook, a RAG platform in Python using FAISS and LangChain, and orchestrated autonomous AI Agents for multi-step legal research with prompt engineering, contextual ranking, and LLM summarization—containerized with Docker and exposed via modular FastAPI REST endpoints. Alongside this, I build foundational engineering work like a Go interpreter (lexer, parser, AST, evaluator) and am currently developing a custom C++ ray tracer using strict Test-Driven Development.