Product Engineer

About BottleCap AI

We're building sovereign AI in the heart of Europe — frontier models developed, trained, and deployed here, by a team of top AI/ML engineers.

BottleCap AI was founded by Tomáš Mikolov — creator of word2vec and one of the most-cited researchers in modern AI — and Jaroslav Beck, founder of Beat Games (Beat Saber). We're backed by 20VC and the founders of Stripe, Hugging Face, and Lovable.

We're a team of 15 in the heart of Prague. Small enough that everything you ship is yours; backed well enough that you're building for the long run.

The role

You'll be one of the first product engineers, owning products end-to-end: from a whiteboard idea through MVP to production and iteration. You'll work directly with our researchers — no PM layer, no handoffs, no committee. You bring engineering reality into research discussions; they bring frontier capability into your products.

Priorities shift fast here. The job rewards tolerance for ambiguity, taste in what to build, and the judgment to know when to scale something and when to keep it simple.

What you'll actually do

• Own product surfaces end-to-end: scope, build, ship, measure, iterate
• Turn research prototypes into products real users rely on
• Make the technical calls — architecture, priorities, what to cut
• Prototype fast, test hypotheses against real usage data, kill what doesn't work
• Decide, with the founders, what's worth building next

What you'd work on first

• Help deliver our improved frontier models to millions — starting with ThinkingCap-Qwen3.6-27B
• Help make AI usage safe with AI Scan
• …and more in the pipeline we'd love to show you at a whiteboard

You might be a good fit if you

• Have shipped a product from zero to one with real ownership — as a founding/early engineer, or with equivalent scope somewhere bigger
• Are a strong full-stack engineer (production frontend + backend)
• Have built with LLM APIs, prompting, or agent frameworks hands-on
• Have shipped to external users and iterated on their feedback
• Think in user value, not tickets

Nice to have: product design instincts good enough to ship a clean UI without a designer; experience with evals or quality harnesses for ML systems; comfort with lightweight data analysis (SQL, notebooks, metrics).

We care what you've shipped, not where you studied. No degree required. If you don't tick every box but this role sounds like you, apply anyway.

What we offer

• Top-of-market compensation — we'll name specific numbers on the first call, before you invest any time in the process
• Contractor (IČO) or employment (HPP) — your choice; we'll walk you through the trade-offs transparently
• Direct access to frontier models and the researchers building them
• Your work in production within weeks, not quarters — no bureaucracy, decisions in days
• A seat at a 15-person company founded by people who've done it before

Why on-site: we ship daily as one room, and the research↔product loop works because it happens at a whiteboard, not in a ticket queue. If that's not for you, this isn't the right role — we'd rather say it now.

Process

• 30-min call with a founder — mutual fit, your questions answered, compensation named
• Technical deep-dive on something you've shipped (no leetcode)
• Paid work sample with the team
• Offer — whole process under 2 weeks