// Machine view — the DOM of this page as a parser reads it
Oleksandr Semishan — AI integrator: I put LLMs to work inside the systems a business already runs
20+ years in IT, from corporations to startups (full work history on LinkedIn). I build AI systems that run in production: CRMs, forecasting, person-data aggregation. I find the real requirement, wire it into the systems you already run, and verify the result with a harness I wrote.
A network of clubs runs mailings, reports, trainer pay and online booking in one system on top of the software it already used.
Financial requests reach the right approvers by flexible rules kept as data: who submits and who approves is configured, not coded. Append-only audit log at the database level.
A deterministic forecasting engine: no language model touches the numbers. It compares teams and predicts the effect of roster changes with a reproducible Monte Carlo simulation.
Collects a person's information from official public and private sources over API. One failing source never blocks the others, and the reason for the skip is logged.
Took over an archive with no backend tests and no /health endpoint. Delivered tests, a run mode and a health check, then moved the production system to a new server with a rehearsed night cutover and rollback. My contribution: 211 of 620 commits.
An AI résumé and job-search workspace: résumé from chat or upload, tailoring to each posting with deterministic skill-gap analysis, job scrapers, Telegram alerts. I found my current role with it.
A language-learning service built around words and listening: study cards plus audio from YouTube and live radio, cut into segments and transcribed. Used every day by me and my family.
A web service that turns a questionnaire about a child into a personalised illustrated and voiced story shown in a page-flip reader.
Playground: price and stock monitor, EXIF cleaner and editor, tetris, maze — small things built to try an idea.
Find the real requirement, wire the model into the systems you already run (POS, booking, CRM, messengers), keep calculations deterministic where they must be.
Numbers computed by code, with an honest range instead of a single figure; dashboards your team actually opens.
Read the process, the code and the data; find where it breaks; leave a written list of fixes ordered by cost.
Short project? $85/hour, up to 20 hours.
How to hire me →Full page: where I fit and how we can work →
Hiring for a short project? Terms, process and comparison with a consultant and an agency →
It goes straight to my Telegram. I reply to the address you leave.
Where and when: remote, from Sofia (UTC+2/+3); clients in the US, Canada and Europe. My working hours overlap with both Europe and the US. $85/hour for short engagements up to 20 hours; longer engagements by arrangement. The scope is agreed after a short call.