About
One engineer, one standard and a specific opinion about where models belong.
Code and Proof is Faizan Hasnaat. Five years building production software, six client products live and carrying paying customers, plus a set of tools built to solve problems that kept recurring across client work.
One person is small, and that is stated plainly rather than dressed up. I do not run ten projects at once. The alternative — implying a department that does not exist — is the single fastest way to lose a technical buyer’s trust, and they always find out on the first call.
What you get is an engineer who wrote the code, on the call, answering the question directly.
Checkable
Why this and not general development
The obvious business to start was a general MVP shop — auth, Stripe, CRUD, deploy, four to eight weeks. It is a real market and plenty of people make a living in it.
It was rejected for one reason that matters: that work is getting cheaper every month as coding agents get better at it, and price competition in that lane only moves in one direction.
Going back through six months of shipped work surfaced something more defensible. The same four commitments appeared in every project, across two clients and three products, without anyone planning them as a philosophy. They were just how the work got done.
Those four are the business now. Everyone can call a model. Far fewer people have shipped the system around one and watched it survive contact with real users.
The four
- 01Deterministic where it matters
- 02Every output traces to its evidence
- 03A human gate before anything irreversible
- 04It heals itself
Range, honestly stated
The wedge is AI reliability, but the underlying work is full-stack and the range is real. What is listed alongside is what has shipped and is running, not a list of things I have read about.
That breadth matters less than it sounds. It is here because “can you also do the backend” comes up on most first calls, and the answer is yes — but it is not why you would hire me.
You would hire me because a system you shipped is confidently wrong somewhere and you need someone who has fixed that before.
The stack, in practice
- Web
- TypeScript, Next.js
- Services
- Python, Django, Celery
- Data
- Postgres, Supabase
- Infrastructure
- Cloudflare Workers, Containers, Durable Objects, queues and cron
- Testing
- Playwright, Storybook — where the front end justifies them
Where this is going
Code and Proof is one engineer by choice rather than by circumstance, working to the same four commitments on projects where being wrong is expensive. Staying this size is the point, not a stage to grow out of.
If you would rather work with an established shop of thirty people, that is a completely reasonable preference and there are good ones. If you would rather have the engineer who wrote it explain it to you, that is what this is.
Who you talk to
Faizan Hasnaat
Engineer
First call, every call. No account manager in between, and no handoff to someone who has not read the code.
codeandproof@gmail.com →