RL Environments Research Engineer

We are The Codest - International Tech Software Company with tech hubs in Poland delivering global IT solutions and projects. Our expertise centers on web development, cloud engineering, DevOps and quality. We have developed our own product, Yieldbird, honored as a laureate of the Top25 Deloitte awards. Our mission is to help tech companies build impactful products and scale their IT teams by boosting IT delivery performance.

• Experience with PyTorch or JAX at the framework level (not just importing a model).
• Familiarity with RL concepts: reward functions, environment design, training loops, evaluation.
• Ability to read ML papers and implement them; reproducing or extending research results is essential.
• Production Python skills: Docker, git, clean code, reproducible environments; notebooks-only experience is insufficient.
• Exposure to any of: model training/finetuning, inference optimization, CUDA/Triton kernels, distributed training, model internals (attention, KV caches, tokenizers).

Nice to have:

• Publications or competitive programming background.
• Experience with MuJoCo, game environments, or simulation frameworks.
• Scientific computing experience (Rust, C++, numerical methods).

Profiles that don't fit:

• Web/backend engineers with AI experience limited to calling LLM APIs, building RAG pipelines, or prompt engineering.
• Data engineers or data scientists working mainly in notebooks and dashboards.
• DevOps/infra engineers without ML depth.

• Design and build MLE/SWE environments and diverse tasks.
• Target a specified language model and satisfy the required difficulty distribution.

Only candidates who have trained a model from scratch or built something where a model learns from an environment are suitable. Profiles limited to web/backend AI experience with LLM APIs, data engineers/scientists working in notebooks, and DevOps/infra engineers without ML depth are excluded.