Building end-to-end ML systems — from credit risk models to agentic RAG pipelines.
Based in Pune, India. Growing focus on quant & fintech applications of ML.
B.E. graduate in Artificial Intelligence & Machine Learning (SPPU, 2026). I like building things that go past a notebook — deployed APIs, working pipelines, and models with real evaluation metrics behind them, not just accuracy screenshots.
GenAI / LLM tooling:
Credit Risk ML API End-to-end credit default risk pipeline — RandomForestClassifier + StandardScaler, benchmarked against a logistic regression baseline. 89% accuracy · F1 0.89 · ROC-AUC 0.96. SHAP explainability, served via FastAPI, Dockerized, with GitHub Actions CI/CD.
SPPU Student AI RAG Assistant — live demo Agentic RAG assistant for students — a LangGraph tool-calling agent over a ChromaDB vector store (BAAI/bge-m3 embeddings + reranker), served through FastAPI with a React frontend, deployed on Hugging Face Spaces + Vercel.
SQL Data Warehouse Project Bronze/Silver/Gold layered ETL pipeline in SQL Server — star schema design, surrogate keys, stored procedures.
Vendor Due Diligence Tool Automates vendor verification by cross-referencing India's MCA registry for ownership, directorship and affiliations, and generates colour-coded compliance reports.
- Sharpening SQL and applied ML for Data Scientist / Data Analyst interviews
- Building toward a Forward Deployed Engineer path, starting with hands-on production ML work