The questions founders actually ask.
Everyone says they do AI now. What’s different here?
Most AI portfolios are demos. I own the architecture of a live generative AI image platform and I’m the on call engineer for its inference service. I designed its async generation path, shipped a Claude tool use agent loop into a desktop product in private beta, and built 9 internal AI tools now in production use: 4 Claude Code skills and 5 MCP servers (GCP, BigQuery, Firestore, Slack, Gmail).
I’m in the US or Europe. How does that work?
I work from Noida, India, on IST (UTC+5:30), with 3+ hours of live overlap with Europe and US East every working day. Day to day it’s async: written updates and a build link instead of standups, plus one scheduled call a week if you want one.
What does the first week look like?
A call to pin down the scope, then a written scope with what ships first, what it costs and when, payment terms included. Repo access from day one. If the scope doesn’t hold together, I’ll say so on that call instead of taking the project.
What if I only need part of this?
Then that’s the project: one AI feature, a React Native build on an API you already have, or a second opinion on an architecture that’s already running. Say what you need in the form and I’ll scope to it.
You run products at Tryco and work full time. Where does my build fit?
In 20 protected hours a week, evenings and weekends IST, and one freelance project at a time. The Appy Pie pager covers Pixazo.ai, never your build. If an incident takes one of those evenings, I move the hours within the same week rather than the milestone, and that week’s written update says so.
Who owns the code, and what happens after launch?
You do, from the first commit: code, repository, infrastructure and accounts sit in your organisation, not mine. For 30 days after the final build I fix anything that breaks, at no charge. After that, a retainer only if you want one.
Can I see the work?
Try The Menu and Trvltrunk have public sites you can open now. Their staff consoles sit behind logins, so on a call I’ll screen share the repos, the CI runs, and what broke in production and how I found it. Rosteva is in private beta and I can demo it. References on request.
Do you work with my stack?
Most of this work is the same shape underneath. What I use:
- AI
- Claude Code, Anthropic API, MCP, agentic workflows, prompt engineering, diffusion inference
- Front end
- TypeScript, React, Next.js, React Native, Redux Toolkit, RTK Query, Tailwind
- Back end
- Node.js, NestJS, Express, FastAPI, Prisma, PostgreSQL, Firestore, Redis, BigQuery
- Cloud
- GCP Cloud Run, Cloud Functions, AWS Lambda, Cognito, Docker, Playwright, Nginx, CI/CD
Tell me what you’re building.
I reply in under 24 hours with a few sharp questions and a suggested next step.
30 days of fixes after the final build, at no charge.