
Systems engineer in Bengaluru
I turn language models into software you can trust: right, fast and cheap, all at once.
I work on LLM agents, retrieval and evaluation, and on the pipelines that take them from a notebook to production.
Selected work
Agentic test failure triage
An agent that reads the logs of a failed test run, decides whether the product code or the test automation is at fault, and links the failure to existing tickets and test cases.
Python, LangGraph, Deep Agents, Drain3, OpenAI
YouTube RAG chatbot
Paste a YouTube link and ask questions about the video. Answers are retrieved from the transcript.
Python, LangChain, Pinecone, FAISS, OpenAI, Streamlit
Content Catalyst
Upload a video and get AI-generated summaries, highlights and edits without any editing skills.
Next.js, TypeScript, Tailwind CSS, Cloudinary, Clerk, Docker
Writing
Recommendations




Piyush Kumar
Team Lead, BrowserStack
I have known Prateek from the first year of his B.Tech and have mentored him in his development journey. He has been very fast at understanding complex system logic. What sets him apart is his curiosity and passion for coding. He is flexible across machine learning, deep learning, NLP, computer vision and web development, which makes him an excellent candidate for any project.