Software Engineer — Frontier AI Systems (Data & Evals)

About the role

We are a boutique engineering consultancy and data lab that helps leading AI model providers tackle complex post‑training challenges. As a Software Engineer you will turn open‑ended research problems into reliable systems, explore unfamiliar codebases and own projects from investigation through delivery.

Key responsibilities

• Build AI evaluation environments, including sandboxed OS setups, Docker‑based tests and mock APIs to assess agent performance on real‑world tasks.
• Engineer synthetic data pipelines that generate, filter, transform and validate high‑quality datasets at scale.
• Create human‑in‑the‑loop tools and annotation interfaces for domain experts to review complex data efficiently.
• Translate ambiguous research requests into concrete technical approaches, experiments and production‑ready pipelines.
• Investigate unfamiliar problems by reading papers, exploring open‑source systems, debugging logs and testing practical solutions.

Required profile

• Proven experience building and shipping complex production software.
• Strong software engineering fundamentals and ability to work across languages and frameworks.
• Comfort navigating partially documented codebases and delivering useful changes quickly.
• Ownership mindset: break down goals, identify next steps, communicate challenges and drive work to completion.
• Ability to thrive with evolving requirements and experimental projects.

Required skills

• Docker
• Linux environments
• Asynchronous processing and concurrency
• LLM API integration (OpenAI, Anthropic) with rate‑limit and cost management
• Experience with LangChain, LlamaIndex or DSPy (optional but advantageous)