Headline

AI Systems Engineer @ TRDDC | Knight-Leetcode(2070) | DSA | Gen AI | Agentic AI | Enterprise Digital Twins

About

I build production-grade AI systems that solve real enterprise problems — not demos, but deployable systems with measurable impact. At TRDDC (Tata's R&D division), I work at the intersection of LLMs, retrieval systems, and agentic workflows: → Built a multimodal RAG pipeline (text, audio, video) that processes enterprise data at scale, cutting query latency by 40% → Designed an LLM routing system across GPT-4, Llama-3, and open-source models — reducing inference costs by 35% without quality loss → Led agentic AI workflows for healthcare using AutoGen and CrewAI, accelerating prototype delivery → As an intern, shipped a full-stack RAG document QA system processing 3,000+ pages/day with 92% top-3 retrieval accuracy I care about making AI systems efficient, cost-aware, and actually useful in production — not just technically impressive. Outside of work, I'm a competitive programmer (LeetCode Knight, Top 5% globally, 800+ problems solved) and a deep believer that strong fundamentals make better AI engineers. 📍 Pune | Open to: AI Engineering, LLM Systems, AI Developer, Python Developer roles.

Location

Pune City

Work experience

SDE-Sytem Engineer at Tata Research Development and Design Centre (TRDDC), Aug 2024 - Present · 2 yrs 1 mo, Pune District, Maharashtra, India, - Led end-to-end development of a multimodal RAG system supporting audio, video, and text — enabling voice and text-based Q&A over unstructured enterprise data. - Optimized vector retrieval using FAISS and ChromaDB, reducing query latency by 40% and improving response relevance at scale. - Designed an intelligent LLM routing system across GPT-4, Llama-3, and open-source models — cutting inference costs by 35% while maintaining response quality. (Awarded: Special Initiative Award) - Led agentic AI workflows for healthcare use cases using AutoGen and CrewAI, accelerating prototype delivery timelines. - Mentored 2 interns, conducting code reviews and guiding system design decisions. Tech Stack: Python · FastAPI · FAISS · ChromaDB · LangChain · LangGraph · AutoGen · CrewAI · OpenAI · Groq · HuggingFace

Research Intern at Tata Research Development and Design Centre (TRDDC), Jun 2023 - Nov 2023 · 6 mos, Pune District, Maharashtra, India, - Built and deployed a full-stack RAG-based document QA system processing 3,000+ pages/day for enterprise use cases. - Benchmarked Llama-2, GPT-3.5, and custom embeddings — achieving 92% top-3 retrieval accuracy in internal evaluations. - Developed a React + Flask application enabling scalable document ingestion and querying workflows used by internal teams. - Presented system architecture and findings to stakeholders and clients; authored end-to-end technical documentation. Tech Stack: Python · Flask · React.js · LangChain · HuggingFace · Vector Databases

Education

Bachelor of Technology - BTech, Computer Science Engineering, Indian institute of information technology Nagpur