Software engineer and AI researcher. I recently completed an R&D internship at IIT Kharagpur, working on NLP and information retrieval with Prof. Koustav Rudra, where I built a 75K+ item QA and multiple-choice dataset from cryptocurrency video transcripts. I continue to coordinate the cryptocurrency IR track at FIRE 2026, organised by IIT Kharagpur.
Alongside research, I have four years of production engineering experience (2022–2026) — event-driven data pipelines, PostgreSQL performance work, and REST/GraphQL APIs for a US-based fintech. Final-year CSE undergraduate, largely self-taught.
Interests: responsible AI, NLP, information retrieval, computer vision.
Selected work
- When Better Eyes Lead to Blindness — a diagnostic study of the information bottleneck in CNN-LSTM image captioning. IJCA, 2025
- Escaping the KV-Cache Bottleneck: vLLM vs HuggingFace on an H100 — engineering notes
- nuqs-presets — production-ready React hooks for URL state
- serverless-wiki-search — in-browser full-text search with SQLite WASM
Elsewhere Website · Google Scholar · ORCID · Semantic Scholar · LinkedIn
Open to research and engineering opportunities.