Models and systems that
actually run in production.
One engineer, end to end. I build ML and backend systems that hold up under real load, then hand over the repository, the runbook, and the dashboards you need to operate them without me.
Six areas of work. One person accountable for all of them.
I take a problem from scope through to deployment. No handoffs between people, no one waiting on someone else's ticket.
Applied AI/ML
Fine-tuned, quantized models on hardware you control, instead of a growing API bill.
Computer Vision
Detection and tracking on live camera streams, tuned to the latency you need.
Backend & Frontend
Python backend and the frontend that ships with it, both built and maintained end to end.
DevOps & Infrastructure
Containers, clusters, and CI/CD with tracing, so you see problems before your users do.
Embedded Systems
Software for hardware that runs unattended: sensors, controllers, edge inference included.
Automation
Manual work turned into a scheduled job that reports success and alerts on failure.
02 / STACK
Tools I use daily and would defend in review. Anything outside this list, I will tell you before the project starts.
03 / PROCESS
Durations are typical for a single well-scoped project, not a promise. Anything larger gets split into phases of this shape.
Scope
I agree with you what is being built, what it must do, and what is out of scope.
Build
Weekly working builds, so progress is something you can open rather than read about.
Test
Automated tests, load and edge cases, and a pass on the failure paths.
Handover
Repository, deployment, and runbook, plus a walkthrough of how to operate it.
Tell me what needs to run.
Send a short description of the problem and any constraints you already know: deadline, budget range, hardware, existing code. If it is not something I can do well, I will say so.