Building machine learning systems, RAG pipelines, and agentic workflows.
Final-year B.Tech CSE
Manipal University Jaipur
ML / GenAI roles
starting 2026
IBM RAG & Agentic AI
Professional Certificate
A handful of projects I'm proud of — production ML systems, agentic workflows, and a multimodal accessibility app. Each shipped, measured, and learned from.
End-to-end system pairing XGBoost price prediction with a RAG-powered Q&A layer over property listings. ChromaDB for retrieval, GPT-4o for synthesis.
Real-time accessibility tool for visually impaired users. Voice-activated camera capture with GPT-4o Vision, built natively in Flutter with on-device hotword detection.
LangGraph-powered conversational agent. Intent detection, FAISS retrieval, and tool execution stitched into multi-turn state for automated lead capture.
Modular pipeline from preprocessing to deployment. StandardScaler, Random Forest training, model serialization, and a Flask service for real-time inference.
More on github
I build machine learning systems that ship — with measurable impact and honest evaluation.
I'm a final-year Computer Science student at Manipal University Jaipur. Most of my work sits at the intersection of LLMs, RAG pipelines, and classical machine learning1 — from a Real Estate Intelligence Platform combining XGBoost with GPT-4o, to a real-time multimodal accessibility assistant for visually impaired users built in Flutter.
I care about measurable impact — R² scores, latency budgets, faithfulness metrics2 — and shipping things people actually use. I'd rather have a small system that works in production than a clever one that lives in a notebook.
Recently completed IBM's RAG and Agentic AI Professional Certificate (8-course series). Currently exploring agentic workflows with LangGraph and looking for full-time roles starting 2026.
¹ Generative AI sits on top of classical ML, not in place of it.
² If you can't measure it, you're guessing — and guessing scales badly.
An evolving collection — what I reach for daily, what I'm learning, what I trust in production.
IBM · 8-course series · 2026
Open to ML / GenAI engineering roles, interesting collaborations, and quiet conversations about agentic systems.