ML Engineer

Role overview A remote-first, full-time role on a Reasoning and Agents team, designing and shipping advanced reasoning architectures for large language models. The work centres on multi-path analysis, self-verification, and pushing models toward domain-grade reliability suitable for high-stakes use.

Responsibilities - Partner with the founding team to shape and execute the Reasoning and Agents roadmap. - Translate ideas from current research into production systems that hold up under real workloads. - Take ownership of a substantial portion of the stack, from initial design through to live operation. - Prototype, evaluate, and harden reasoning techniques such as self-consistency, verification, and structured tool use. - Contribute to evaluation infrastructure that measures reliability on domain-specific tasks.

Requirements - Hands-on experience with PyTorch or JAX and modern LLM training or inference frameworks. - Strong background in neural-symbolic reasoning, formal methods, or related areas. - Demonstrated track record of deploying machine learning systems that other people depend on. - Comfort moving between research papers, prototypes, and production code.

Nice to have - Familiarity with agent orchestration patterns, planning loops, or constrained decoding. - Experience with evaluation harnesses, benchmarks, and reliability metrics. - Publications or open-source contributions in reasoning or verification.

Benefits and work setup - Remote-first with a London hub available. - Three-stage interview process: introductory call, technical deep dive, and founder interview focused on vision and impact. - An equal opportunity employer with a stated commitment to a diverse and inclusive environment.