I don't take a job for hours on a bill.
Some get paid when projects ship. I get paid when they live.
I take the job when I understand why it matters and can deliver a result. Usually contracts from three months, iterations every 1–2 weeks.
Sometimes a focused fix: the product is stuck, the backend cannot hold, a feature has not shipped in three months. I come in, find the problem, leave code the team can keep.
Not a fit if
- The result cannot be counted: «it feels more modern» is not a metric
- Nobody inside owns the result
- You need someone to close tickets without context
- The budget is sized for a junior or offshore team
- You need your own model trained from scratch
- You need results 'yesterday' without a real conversation first
Work formats
Diagnostic
I look at what gets done by hand, where a model fits and where it does not. A written analysis with a metric and an estimate. «Do not build this» is a legitimate outcome and it is included.
- Duration
- 5 working days
- Price
- 90 000 ₽
- Outcome
- A document of several pages
- Next
- Fee is deducted from the project
AI inside your backend
A working pipeline in your system: queue, retries, validation of model output, human fallback, monitoring and a cost meter. The success metric is written down before the start.
- Duration
- 4–8 weeks
- Price
- From 350 000 ₽, fixed
- Stack
- Python · FastAPI · PostgreSQL · Redis
- Outcome
- Code, deploy, metric dashboard
Product Rescue
Product stuck: fragile MVP, backend breaking under load, feature not shipping for three months. I come in fast, find the real problem, fix it, and leave clean code behind.
- Duration
- 2–6 weeks
- Rhythm
- Intensive, daily status
- Stack
- Situation-dependent
- Outcome
- Fixed, measurable
Support and long-term contract
The system keeps living: I watch quality and token cost, fix drift, extend it. I get deep into the domain and take part in architecture decisions.
- Duration
- 3+ months
- Price
- From 250 000 ₽ per month
- Rhythm
- 1–2 week iterations
- Format
- Async-first, calls when needed
Questions & answers
How fast do you respond?
On working days, within a few hours. Writing first: calls when needed, I don't disappear for a day.
What if the task changes mid-project?
That's normal, it's part of any living project. We note the change, update the scope, move on. I don't cling to the original agreement if it no longer makes sense.
Do you work in a team or only solo?
In a team, yes: code reviews, shared architecture decisions, onboarding colleagues. Team lead is a separate conversation.
Do you take frontend tasks?
No. Backend, API, infrastructure: yes. If you need fullstack, I'll say so and won't take it on.
Do you give pricing upfront?
The diagnostic costs 90 000 ₽ — that price is on the site and does not move. I quote the implementation after it: before I understand the scope, a number would not be fair to either side.
Do you work with NDAs and contracts?
Yes, standard practice. I sign NDAs and work under a contract. I have templates, we can discuss.
What does the start of work look like?
Short call (30 min) → one-page scope doc → repo and environment access → first tasks in the first week. No bureaucracy.
Got a task. Write.
I reply within a day. No pitch: we'll see if we fit.