Open to new opportunities
I turn LLMs from demos into company infrastructure.
Nineteen years in IT: from database administration to engineering leadership and back to building. Today that means the platform layer that makes AI usable inside a company — and it means shipping it, not demoing it.
Remote-friendly · on-site ready · full-time priority
AI / Platform Engineer. I started in 2007 as a database administrator, spent a decade building high-load and ad-tech systems, led engineering teams of up to thirty people, and came back to hands-on work in the place I find most interesting: the platform layer that decides whether AI actually gets used inside a company.
Over the past six months I shipped about twenty internal services to production — support analytics on RAG, a gateway in front of the model provider, a secrets platform, a marketplace of shared AI tooling and documentation that maintains itself. They are used daily by engineering, QA, product and HR.
The through-line across nineteen years is the same: I like the boundary where engineering meets operation. Anyone can demo a model, ship an MVP or promise a migration. The work that matters is SSO, audit trails, migrations, CI gates, monitoring and documentation that does not rot — the things that decide whether it is all still running six months later.
A prototype that cannot be operated is not finished. SSO, audit, migrations and a rollback path are part of the feature, not a follow-up ticket.
Making automated tests mandatory and rebuilding deployment did more for stability than any individual hotfix ever did. I would rather change how a team ships than firefight what it shipped.
Every service assumes its dependencies may be missing: counters fall back to memory, file storage falls back to the database, an exhausted key drops out of the pool by itself.
Knowledge that lives only in someone's head is an outage waiting for a vacation. A document maintained by hand is wrong within a month.
Feb 2026 — Sep 2026 · Batumi, Georgia
Internal AI and platform layer: RAG-based support analytics, a secrets platform, a marketplace of shared AI tooling, an LLM gateway and self-maintaining documentation. Standardised service delivery onto shared CI pipelines and a single convention for configuration and secrets.
Dec 2024 — Feb 2026 · Batumi, Georgia
Jan 2024 — Oct 2024 · Madrid, Spain
Led the supply-side platform project and assembled the team around it, with a mandate to reach results in a short window.
Mar 2022 — Nov 2023 · Moscow
Joined as the advertising platform was being formed and took it from scratch to a revenue-generating product within a year. Owned all fourteen engineers on the project — backend, frontend, QA, later mobile — and split the domain into three teams once the MVP shipped. Designed the services and subsystems behind a classic RTB stack.
Feb 2019 — Mar 2022 · Moscow
Joined as a senior developer and grew into technical ownership of a project with more than thirty people. Made automated tests a mandatory part of development, planned the retirement of legacy subsystems, hired and split the group into three teams on the way to microservices, and rebuilt deployment with real monitoring and alerting.
Mar 2014 — Oct 2018 · Moscow
Started in Lua and moved to the ETL team to scale an existing data-processing system. Designed and built a microservice-based subsystem for pre-release automated testing.
Jul 2012 — Mar 2014 · Moscow
Technical lead for a publishing house's digital library: backend and frontend, server procurement and setup, monitoring and alerting, new subsystem design, and quality control over external contractors.
May 2011 — Jul 2012 · Moscow
Automated content updates for satellite sites — faster, more stable and with verification of results — and optimised the database and its queries.
Sep 2008 — May 2011 · Moscow
Database infrastructure and stored routines: schemas from scratch, performance analysis, storage optimisation and bottleneck elimination.
Nov 2007 — Sep 2008 · Moscow
Where it started: relational database architecture, performance tuning, stored procedures, clustering and replication.
Bachelor's Degree, Computational Science
Inspiring and Motivating Individuals