Building things — some serious, some just fun.
A mix of protocols, apps, research, and whatever felt worth making.
I'm Tanay Kashyap. By day I'm an AI-enabled software engineer; alongside that I'm finishing a Master's in AI & ML — and the rest of the time goes into the things below. Some ship under Baniloo Labs — a small org I started for the open protocols and tools: a medical records protocol, a claim validation network, and Postmortem, an AI ops watcher that lives in your terminal. The rest I build solo, under my own name — a federated-learning dissertation, a task tracker I actually use, a gym app, a World Cup fan experience. The range is real.
Some of these are designed from first principles. Some started because the problem annoyed me. Either way, they're documented here as they get built — split below into Baniloo Labs and the solo work.
Ideas
All ideas →AI-powered ops intelligence that lives in your terminal. It watches your deploys, git, and logs — and when something breaks, it correlates the events and tells you why, using whatever AI you already have. Local-first, no SaaS.
An open protocol for validating claims. A reputation-weighted validator network produces permanent, tamper-proof verdicts on falsifiable assertions — without storing content or owning the truth.
A protocol for patient-controlled medical records — portable, interoperable, and built without requiring hospitals to cooperate first.
A real-time fan reaction app for FIFA World Cup 2026. Pick your team, react as the match unfolds, watch a live world map show where everyone is feeling it.
A pay-per-use gym and fitness venue app. Walk in, scan a QR, train, scan out. No memberships. A full session lifecycle with wallet tokens, venue discovery, and a gym owner portal.
A personal life management and task tracker. Kanban, calendar, a Today view that tracks momentum and drift, recurring tasks, multiuser Spaces, and Streams.
Master's dissertation in AI & ML. Federated learning for Long-Term Care facilities — predicting skilled nursing needs without ever moving patient data off-premises.
Exploring how the world will move from video-based learning to AI-influenced learning maps.