3 hours ago
Cambridge, United KingdomSenior
Responsibilities
- Own fulfilment-ranking pipelines from training orchestration and quality gates through deployment and online validation.
- Build and operate an offline evaluation platform using LLM-as-judge harnesses and deterministic quality checks.
- Develop the LLM agent, including tool routing, retrieval, guardrails, and answer caching.
- Design caching and observability systems for latency, quality, tracing, reliability, and per-request cost.
- Improve distributed-system reliability and lead diagnosis of difficult cross-service failures.
- Drive accountable AI-assisted engineering practices and collaborate across machine-learning, product, data, and platform teams.
Requirements
- Strong production software-engineering experience designing, testing, operating, and debugging distributed services.
- Meaningful ownership of ML infrastructure or production model delivery with evidence of operational or adoption impact.
- Practical understanding of training, evaluation, versioning, deployment, online validation, monitoring, and safe rollback across the ML lifecycle.
- Experience designing for latency, resilience, observability, and cost in production systems.
- Experience with LLMs, embeddings, semantic search, retrieval, or agent systems, including quality evaluation and failure-mode control.
- Experience with caching systems and low-latency service trade-offs.
- Commitment to automation, CI/CD, code quality, and evidence-led engineering decisions.
- Collaborative communication skills and ability to explain technical trade-offs across disciplines.
- Bachelor-level degree in Computer Science, Electrical Engineering, or a related field, or equivalent practical experience.
- Experience with AWS or GCP, Kubernetes, SQL warehouses, Trino, Presto, Spark, Kafka or other streaming systems, Go, or Python is useful; Node.js experience is a plus.
Benefits
- Benefits may include healthcare, dental and vision coverage, life, accident and disability insurance, commuter benefits, retirement options, mental-health and financial-wellness support, and time off subject to local policies.
- The role generally works in the office Monday through Thursday, with Fridays generally flexible for remote work; some roles or locations may require five days per week onsite.
Categories
About Roku
Roku builds streaming players, Roku-branded TVs and audio gear, and the Roku OS licensed to TV manufacturers. Its platform supports ad-supported and subscription streaming, including The Roku Channel and Roku Originals, and underpins a significant advertising business. Founded in 2002 and headquartered in San Jose, it is a public company trading on NASDAQ under the ticker ROKU.