Senior Software Engineer, GenAI Platform

About the role

Deliveroo's GenAI Platform team, within Machine Learning Platform, builds the shared infrastructure that lets DoorDash, Wolt and Deliveroo teams safely ship GenAI products, agents, automation and personalisation. The team runs frontier open-weight LLMs and VLMs itself (real-time GPU serving, high-throughput batch inference and fine-tuning on autoscaling GPUs) and owns the LLM Gateway, Agent Gateway, evals infrastructure, guardrails and cost attribution. You will lead the design and architecture of the open-weights model platform spanning inference and fine-tuning, set technical direction and mentor engineers.

Key responsibilities

• Lead the design of infrastructure that moves GenAI ideas from prototype to production.
• Own and evolve the open-weights serving stack (real-time GPU endpoints, batch inference, fine-tuning) alongside the LLM Gateway, Agent Gateway, evals, guardrails and cost attribution.
• Architect scalable, high-performance systems for model serving, batch inference, GPU autoscaling and fine-tuning.
• Push the cost and latency frontier of GPU inference, improving batch job speed and reducing inference costs.
• Build platforms that support rapid experimentation while meeting production standards for latency, scale, monitoring, SLOs and operational excellence.
• Partner with ML engineers, product engineers and data scientists across DoorDash, Wolt and Deliveroo.
• Set technical direction for the centralised GenAI platform, including RLHF/RLVR, agent optimisation and other post-training techniques.

Required profile

• BSc, MSc or PhD in Computer Science or equivalent.
• 5+ years of industry experience in software engineering.
• Track record of designing and owning production services, APIs, data pipelines or ML infrastructure at scale.
• Demonstrated technical leadership: leading design, mentoring engineers and turning customer use cases into reusable platform capabilities.

Required skills

• Deep backend engineering fundamentals, especially Python and distributed systems.
• Hands-on experience with LLM inference and/or fine-tuning of open-weight models in production (SFT/DPO/LoRA).
• Experience operating systems in production, including observability, debugging, reliability, incident response and performance/cost optimisation.
• Nice to have: vLLM, SGLang, TensorRT-LLM, Kubernetes, AWS/GCP, GPU platforms (e.g. Modal), LLM gateways, eval systems, RAG, vector databases, and AI coding tools such as Claude Code, Codex and Cursor.