Ganesh Epili

Hello, my name is Ganesh. I'm an AI Engineer and NLP Researcher at IRLP Lab, DAIICT, where I build intelligent systems that bridge the gap between cutting-edge research and real-world applications. My focus is on making Large Language Models practical, reliable, and accessible.

My research centers on Retrieval-Augmented Generation and repository-level code understanding. I'm currently working on RepoRAG — a system that overcomes context length limitations of LLMs to enable code generation and debugging across entire repositories. The paper is under review.

I've also built IndicTranse, a voice-to-voice translation platform for Indian languages using a fine-tuned phi-2 2.7B model with real-time ASR and TTS, and IntellectRAG, a multi-source Q&A system powered by Mistral 7B that works across documents, webpages, and YouTube transcripts.

Before research, I worked as a Backend/ML Intern at Jungleworks (Jugnoo), where I built a platform for creating custom chatbots on client data using Node.js and RASA. I was awarded Best Employee for this work. I also developed interactive analytics dashboards and optimized the codebase to reduce graph loading time by 70%.

On this site, I write about AI, NLP, and lessons learned from building systems that go from research papers to production — a practice I plan to continue daily.