Vi bygger systemer som jobber om natten for polske selskaper i dag — og som vil jobbe for selskaper i hele Europa i morgen. Vi leter etter folk som får mer tilfredsstillelse av å jobbe med språkmodeller, prosessautomatisering og ekte produksjons-deployment enn av nok et møte om nok et lysark.
No POCs that end up in a drawer. Every project ships to production for a paying client. Your work is measured by impact, not closed tickets.
Claude 4.7, GPT, Llama, FastAPI, Next.js 15, MCP, pgvector. No struggling with a decade of legacy, but we don't chase every trend either — we pick what works.
Team of up to twenty people. Decisions made the same day, not after three committees. Your idea for a new project can launch on Wednesday.
We have an office in Warsaw, but 80% of work is remote. Flexible hours — we care about outcomes, not attendance. 4 weekends per year of in-person team workshops.
Every role is work on production systems for real clients. No artificial recruitment tasks — during the interview we show you the project you'll be working on in your first week.
Work on the core of our systems — agents, RAG pipelines, Claude integration via MCP. Code you write ships to a client's production within the same week.
Send CV and a short note about yourself. No cover letter — a link to a project you're proud of will do.
SøkSvar innen 5 virkedager.
The link between client and dev team. You talk to the client about their process, design the solution, and run the deployment from zero to production.
CV + a short description (3–5 sentences) of an AI project you led from idea to production.
SøkSvar innen 5 virkedager.
Your work after the product ships to the client — training the team, measuring impact, tuning the system to real work. The difference between a project that 'works' and one that 'sticks'.
CV + one example: what specifically you've automated for yourself or your team with AI tools.
SøkSvar innen 5 virkedager.
Studying computer science, cognitive science, mathematics, or just finished a bootcamp and want to try working with language models? We have a place for you — not to brew coffee, but to do real work on products that ship to clients.
We run internships three times a year: February–April, June–August (student favorite), October–December. Each internship is 3 months at 20–30h/week (negotiable around your schedule).
Each intern gets a mentor (one of our seniors you can call mid-day), a concrete project (not a list of 'rewrite the README' tasks), and a weekly 1:1 with feedback. After 3 months we usually propose continuation — either full-time or help finding work elsewhere if we're not a fit.
Standard intern compensation (B2B or contract of mandate). For full-time after the internship — junior range from the Product Specialist role.
If your university requires a mandatory traineeship — we accept unpaid traineeships of 80–160 hours. We issue a certificate + reference letter.
We accept applications anytime — recruitment for the next cohort starts 6 weeks before launch.
Cognify is growing — we open new roles every quarter. If your experience doesn't fit any role above but you believe you bring something we're missing — send us a message. We read every application.
Skriv do nas