Working things I've built hands-on. Everything here is client-safe.
Role-based operations intelligence platform for a rail container-logistics network — executive, operations, HR, customer, finance, safety, and sustainability modules with GenAI-driven alerts, forecasts, and what-if scenarios over 1,200+ container records. This is the anonymised version of the prototype that moved a ₹10 Cr deal into commercial closure.
🔗 Live demo — demo login: admin / admin
Full-stack predictive-maintenance platform for marine vessel fleets — ML-driven oil degradation forecasting (XGBoost classifier, LightGBM remaining-useful-life regressor, Isolation Forest false-alarm suppression), live sensor feeds, an autonomous retrain agent, crew job cards, and alert-to-recovery workflow. Built with synthetic data for an innovation-showcase setting.
🔗 Live demo — demo credentials pre-filled on the login screen
Started as a product to solve my own problem — forgetting where I keep things at home — and evolved into workplace asset tracking. The interesting product decision: the same core mental model (“every item has a home zone; flag it when it moves”) scales from car keys to industrial equipment. Mobile-first, with spaces & role-based invites (owner / admin / member), check-in/check-out flows, QR scanning, and voice search.
🔗 Live demo — tap ⚡ Demo Home or ⚡ Demo Office to skip onboarding
A “living case brief” for solo Indian advocates and small chambers — one file per matter that keeps itself current: order sheets synced from eCourts/NJDG, uploaded documents and voice notes folded into a chronology, and drafts generated only from the matter's own record. Two design decisions I'm proud of: per-matter “standing rules” (counsel's red lines, enforced on every generated draft) and a visible “grounded in / not used” panel — every draft cites its sources and explicitly refuses to invent case law.
Construction-tech venture I founded (2017): CAD plug-in SaaS for cloud drawing management — 50% efficiency gain for architect workflows. Finalist, Government of Karnataka startup funding programme. Currently exploring an AI/ML-modernised concept.