Senior Python Engineer

VirtusLab is a leading European software consulting and engineering company, home to over 350 EU-based professionals. Their mission is to craft clean code and practical solutions with precision and purpose. The company fosters a dynamic culture rooted in strong engineering, a sense of ownership, and transparency. The role is within a small, startup-like team building a tool for deploying and configuring local AI inference on local devices like NVIDIA Spark, managing model lifecycle and connecting coding agents to locally hosted models.

• Strong Python skills with experience in backend frameworks such as FastAPI or Flask.
• Solid understanding of concurrent and asynchronous programming.
• Ability to rapidly prototype and iterate on emerging ideas.
• Engineering pragmatism: cutting complexity while keeping the system extensible.
• System-level thinking: understanding Docker lifecycle, LLM proxy routing, model management, and agent layer integration.
• Hands-on experience using LLMs daily with good judgment.
• Self-motivation and ability to take full ownership of features end-to-end.
• Fluency in English with good communication skills for a remote-first team.

Nice to have:

• Experience using Docker for local deployment and service orchestration.
• Experience setting up and maintaining an LLM proxy layer such as LiteLLM or OpenRouter.
• Experience with vLLM or ollama.
• Familiarity with the HuggingFace API and model distribution patterns.
• Experience with local GPU inference setups (CUDA, NVIDIA hardware).
• Background working with coding agents (Claude Code, Codex, Aider) and their configurations.

• Build and maintain backend systems using Python and frameworks such as FastAPI or Flask.
• Develop and manage the lifecycle of Docker images that ship AI agents.
• Adapt LiteLLM configuration as agent APIs evolve.
• Integrate with the Hugging Face API for model downloads.
• Implement production essentials such as idempotent endpoints, retries with backoff, and migrations.
• Work on managing model lifecycle (start/stop) and connecting coding agents like Claude Code and Codex to locally hosted models.
• Collaborate in a small team with direct contact with potential users and fast iteration.