Staff Software Engineer
London, UK

About the role

Applied Intelligence Systems (AIS) is part of the Scale Generative AI Platform (SGP), focused on pushing the frontier of what agentic applications can do across diverse enterprise and government use cases. We build the infrastructure and tooling that power agentic AI in production, paired with applied ML research, design, and evaluation to ensure these systems perform reliably at the scale our customers demand. AIS spans multiple workstreams — agent evaluation and oversight, orchestration and tool-use infrastructure, model and systems optimization, and applied research on new agent capabilities. This role owns the context and memory capabilities within AIS, including their correctness, performance, and evaluation. 

We are looking for a Staff Engineer who can own hard technical problems end to end, from the software systems that serve context to agents through to the evaluation choices that determine whether that context is actually useful. This role suits someone comfortable moving between distributed systems engineering and applied ML, because the team is built the same way, with software engineers and ML engineers working the same roadmap rather than two separate tracks.

What you'll do

• Own large, ambiguous problems in context and memory end to end, from design through production, including the backend systems, the retrieval and memory algorithms, and the evaluation that proves they work.
• Architect the core primitives that let agents retrieve, store, and reason over long-running and cross-session context, including knowledge base retrieval, vector stores, and memory strategies.
• Design and maintain the evaluation methodology for memory and retrieval quality, including the rubrics and benchmarks that catch regressions before customers do.
• Set technical standards that other engineers on the team adopt, whether that is an architectural pattern, an eval practice, or an approach to failure handling under partial system failure.
• Partner with SWE and MLE peers on the same roadmap, and coordinate with AIS's other workstreams (orchestration, evaluation and oversight, systems optimisation) where memory and context intersect their scope.
• Debug and resolve the most severe production issues tied to context and memory, including incorrect retrieval, stale or leaked context across tenants, and degraded relevance at scale.

What we look for

• 8+ years of engineering experience, with a multi-year track record owning production systems end to end, not just implementing scoped work handed to you by another team.
• Direct experience with the ML and information retrieval problems underneath modern AI systems, such as embeddings, vector search, retrieval augmented generation, fine-tuning, or agent memory architectures, and the judgement to choose between them for a given problem.
• Comfort owning both sides of the stack. You do not need to be equally deep in both software engineering and applied ML, but you need enough range in each to make good calls without waiting for a specialist.
• Proficiency in Python, and experience with the infrastructure that production ML and agentic systems run on (containers, cloud platforms, and CI/CD).
• Track record of setting technical standards that outlived the project they were built for, and of influencing decisions beyond your own team through design reviews, technical writing, or direct mentorship of software and ML engineers.

Ideally you'd have

• Experience scaling products at hyper growth startups
• Experience with agent orchestration frameworks and multi-agent systems in production.
• Contributions to open-source agentic AI projects.