Things I build for myself.
I build the tools I can't buy: bots, automations and systems that save me time or teach me faster. This is what runs in the background.
Self-improving second brain
A fully-built Obsidian OS that auto-scrapes viral reels, LinkedIn posts and podcasts, distils them into a structured knowledge graph, runs a nightly synthesis pass across every note, and quizzes me with active recall to test what actually stuck.
Research-analyst bots
Auto-bots that do the grunt work of equity and crypto research: screening, summarising and flagging, so I act on signal, not noise.
Financial models & research
Custom-built valuation and DCF models that tear a business apart before I commit a rupee, alongside whatever research tooling the off-the-shelf products don't offer.
The content engine
A pipeline that turns my daily inputs into scripts and posts, the system behind 4M views in 90 days.
Vibecoded, shipped, forgotten
Internal tools, dashboards and weekend experiments I vibecode and ship just to see if I can, most of them without ever telling anyone.
And plenty I do just because.
Not everything has to ship or scale. A lot of what I do is pure curiosity, and it quietly makes everything else better.
Productivity websites
I build small productivity tools and websites just because I enjoy it, scratching my own itch and shipping something useful by the weekend.
Learning AI
I learn AI for fun: agents, automations, prompt engineering, and wire whatever I learn straight back into the things I'm building.
Podcasts & deep dives
I watch podcasts and long-form deep dives constantly, mining them for ideas on markets, building and how people actually think.
Learning, constantly
I learn a lot of things for the sake of it. The compounding curiosity is the point. New rabbit hole every week.