End-to-end website for a real wellness studio: Next.js 16 App Router, TypeScript, Tailwind CSS 4, dynamic pages, and an admin dashboard.
An in-house IT system for India's iconic pressure cooker brand.
15 years of national grain data, forecast and audited.
Synthetic event warehouse, A/B test, full pipeline.
A dual-mode RAG system for querying financial reports, 10-Ks, and strategy decks — runs fully offline for sensitive workloads, or cloud-deployed for demos and non-sensitive use cases.
Hindi, Bengali, Marathi, English — one chat surface.
Text + image inputs, vision-aware responses.
Natural language → executable SQL.
15K transactions, segmented and visualized.
9,000 job posts scraped, skills benchmarked.
I'm Alok Deep — Master's in Computer Applications (Data Science) from Dev Sanskriti Vishwavidyalaya, currently in Bengaluru. My recent stretch has been split between two tracks: building production-style data systems (HDIP, foodgrain forecasting, ShareChat analytics) and shipping AI/NLP applications during my time at NullClass Edutech.
What sets me apart is range. I'm comfortable writing the SQL ETL, training the forecasting model, deploying it behind FastAPI, and styling the React dashboard that visualizes the output. The MERN-stack background means I treat ML as a product, not a notebook.
Outside of data work I freelance on web dev, drink too much filter coffee, and read papers I half-understand.