whosgotch

Location: Kyiv, Ukraine

Remote: Worldwide

Willing to relocate: US (sponsorship required)

Technologies: Python, Go, TypeScript, PostgreSQL, Redis, Kafka, Docker, AWS, FastAPI, Django, MCP

Résumé/CV: Available on request

Email: komovaar@gmail.com

LinkedIn: https://www.linkedin.com/in/komovaar

GitHub: https://github.com/komovaar

Backend and AI infrastructure engineer with 5 years of experience. I build production systems and developer tools, usually owning the whole thing from design to deployment.

At Hitachi, I built the model infrastructure behind an AI coding platform used by 120+ engineers, led our first production MCP server controlling 20+ TV models, and cut a regression suite from 27 hours to 6. I've also worked on payment systems, distributed task processing, and reduced production p95 latency by 65%.

I build open-source tools too. Kolo (https://github.com/komovaar/kolo) brings Google Docs-style collaboration to AI coding agents. Hewitt (https://github.com/hewitt-iso/hewitt) runs agents in isolated Docker containers and Git worktrees.

Looking for backend, AI infrastructure, or developer tools roles. I'd love to join a small team building something technically ambitious, where I can own real problems and help make engineering decisions.

Happy to talk directly with founders and engineers.

Location: Kyiv, Ukraine Remote: Yes Willing to relocate: Yes Technologies: Python, Go, JavaScript, TypeScript, Lua, C, SQL, PostgreSQL, MySQL, MariaDB, MongoDB, Django, FastAPI, Flask, Odoo, Redis, Elasticsearch, FastMCP, Kafka, Docker, Grafana Resume/CV:

Software engineer with 5 years of experience, split between backend systems and AI tooling. At GlobalLogic I built the internal AI coding platform now used as the standard across the org by 120+ engineers - an OpenCode fork with a custom model provider layer, internal service integrations and a rebuilt UI. I also shipped the company's first production MCP server driving physical hardware, giving agents direct control of client TV devices across 20+ models, and cut a global streaming platform's QA suite from 27 to 6 hours, which made nightly regression runs possible for the first time. Alongside that I wrote and delivered the internal AI-engineering curriculum to 60+ engineers.

For more information, please contact email/linkedin

Email: komovaar@gmail.com LinkedIn: https://www.linkedin.com/in/komovaar

I've been building something in this direction - instead of a chat thread, agents are nodes in a graph with explicit dependencies, so you can fan out a few in parallel and see where each one is. The control story is much better than a chat. Not my daily driver for coding yet.

https://github.com/whosgotch/orbital