Projects I've built — from production ML systems and MLOps tooling to side experiments and utilities. Sorted by recency, with featured work highlighted.
⭐ Featured Projects
Best representation of my current skills: ML systems, secure code execution, MLOps, and applied AI.
Government transparency platform covering every Central department and all 28 states + 8 UTs — budgets, expenditure vs budget, CAG audit findings, and a live PIB press-release pipeline behind 15 FastAPI endpoints.
September 2026CLI that scans your hardware, estimates what actually fits, and recommends local LLMs worth running — with VRAM-aware model comparison and Ollama chat built in. Published as localmind-ai on PyPI.
July 2026Full-stack platform (built at Smartivity) managing the end-to-end product design lifecycle — Slack OAuth, per-project channels, timezone-aware reminders, and analytics dashboards, used by 100+ designers and managers.
July 2026AI-driven platform to discover and apply for government welfare schemes using eligibility matching and vernacular support.
June 2026Docker sandbox with AST scanner + YAML policy engine for secure execution of untrusted AI-generated Python code.
May 2026Energy optimization platform using predictive analytics and real-time monitoring to reduce industrial power consumption.
April 20264.2M parameter language model trained from scratch using PyTorch — deep dive into training workflows and architecture.
February 2026End-to-end MLOps prototype: FastAPI backend, Streamlit dashboard, Discord alerts for real-time drift detection.
February 2026Health assistant for rural areas — symptom checker, medicine reminders, and nearby clinic locator with offline-first support.
October 2025📦 Archive & Experiments
Side projects, learning experiments, and utilities — sorted newest first.
🧩 Full source code and more experiments on GitHub — 50+ repositories, 40+ merged PRs.