I build AI-driven products for teams moving through ambiguity: tools that make decisions clearer, workflows simpler, and shipping safer.
My product education started the expensive way. I co-founded a short-video app, took it to 10,000+ downloads in one week, then shut it down ten days later. That experience taught me what no framework could: find the real constraint, ship the smallest useful thing, and learn before certainty becomes expensive.
Today, I bring that lesson to 0–1 product work and hands-on builds across AI agents, developer tools, analytics, and consumer experiments.
| Project | What it does | Explore |
|---|---|---|
| Fumble Machine | Turns past purchases into shareable “what if I invested instead?” receipts using real market data. | Live product · Source |
| TokenBurn | A privacy-first browser extension that estimates AI token use locally and turns it into a visual activity trail. | Source |
| ProdReady | Checks AI-generated web apps for security leaks, SEO gaps, and production-readiness issues before launch. | Live product · Source |
| Agentic Product Manager (WIP) | Planned agent workflow for stress-testing product ideas, checking market signals, and producing an opinionated build sequence. | Source |
| Chaitanya.lol | My product work, experiments, and lessons from building across B2B, B2C, and AI. | Read the story |
- Find the real problem — talk to users, inspect behavior, and separate signal from requests.
- Reduce the bet — turn uncertainty into the smallest test that can change a decision.
- Ship the loop — build, measure, learn, and keep the feedback cycle short.
- Make it reusable — turn hard-won learning into systems, tools, and clearer defaults.
Current focus: AI product strategy, agent workflows, visibility measurement, and trustworthy shipping.
Tools I reach for: TypeScript, Python, Next.js, SQL, product analytics, and whichever model best fits the job.
If you are building something useful at the edge of product and AI, say hello.