Senior Applied AI Engineer

About us

• Software lab - Flyps is a high-tech software house with nearly a decade of experience turning complex technical challenges into production-ready products.
• Clutch 5.0 from satisfied customers!
• Products we helped build were later acquired!
• (recent exits to Cursor, SpaceX, LlamaIndex, PE)
• Work with serial entrepreneurs, SF startups, and established businesses that want to stay ahead of market shifts.
• A highly experienced team, with in-house PhDs and most senior engineers.
• Work at the technological frontier:
• An actively developed internal AI harness with access to leading models.
• Infrastructure for experimenting with local models but don’t worry, we haven’t lost our heads and gone all-in on tokenmaxxing :)

The role

We're looking for a Senior Applied AI Engineer to build AI systems for our clients and for Flyps itself, and to lead other engineers technically. Our projects cover the whole field: classical machine learning, LLM and multimodal applications, agentic systems, image and speech generation.

We build software agent-first: our engineers use coding agents as their main way to write code, and spend their own time on specification, architecture and verification. We don't hire for a framework. Most of our AI work runs on Python and TypeScript, so you should be able to work across the whole stack - backend, frontend, data and deployment. Coding agents make syntax cheap, but you still need to judge whether the result is right.

What you’ll do

• Build systems from idea to production, including backend services, data pipelines, integrations and monitoring.
• Prototype quickly to test an idea before the team commits to a full build.
• Choose the simplest approach that works - sometimes a gradient-boosted model, sometimes an agent, sometimes no AI at all - and compare options by quality, cost, latency and reliability.
• Build evaluations and regression tests, so every change shows whether it actually improved the system.
• Diagnose failures from code, data, logs and execution traces.
• Work with clients from discovery to delivery: scope projects with sales, agree acceptance criteria, and explain decisions, results and limitations.
• Build internal AI tools, including our shared agent tooling, and help the whole company - engineers and non-technical teams - work better with AI.
• Set technical direction, review solutions, and mentor engineers.
• Test new models, tools and methods, and recommend only what holds up on real tasks.

What we’re looking for

• Full-stack engineering: you have shipped production applications end to end, with automated tests, Git and code review.
• Agentic engineering in daily practice: you build with coding agents, keep them on track in a large codebase, and check their output instead of trusting it.
• Agentic systems in production: tool use, multi-step workflows, guardrails, and evaluation of agent behaviour.
• Production experience in at least one more area of AI: classical ML (tabular data, time series), deep learning and computer vision, LLM applications and RAG, or generative media (image, audio, video).
• Solid ML fundamentals: data quality, choice of metrics, test sets, and telling a real improvement from noise.
• Technical leadership: you have made architectural trade-offs, explained them, and mentored other engineers.
• Clear communication with clients and engineers - you can explain a decision and challenge an assumption.
• Comfortable working in a bilingual environment (Polish C2 and English C1).

Nice to have

• Fine-tuning or distilling open-weight models (e.g. LoRA), running local inference.
• PyTorch, scikit-learn, XGBoost / LightGBM.
• MLOps: experiment tracking, model serving, GPU infrastructure.
• Your own agent harnesses, MCP servers or tools for coding agents.
• Image or speech generation pipelines.

What we offer

• 26 000 - 33 000 PLN net + VAT monthly, B2B
• 100% remote setup and flexible hours
• Private medical care, life insurance, and Multisport
• Referral bonuses
• Equipment (MacBook by default)
• Interesting, challenging projects - we only take on the ones we believe in :-)