Platform Engineer (DevOps)

About Us Food waste costs the world over $1 trillion a year. At Winnow, we're on a mission to change that — bringing AI-powered technology into hospitality kitchens to help eliminate waste at scale, with an ambition to save our clients $1 billion annually. We started in London in 2013 with a simple belief: technology could make kitchens smarter and food systems more sustainable. Today that belief has grown into a global team spanning the UK, Romania, Dubai, Shanghai, Singapore, Malaysia, North America, and Mexico, working across 90+ countries and 3,500+ sites. We're proud of the results we help our clients achieve. Names like IKEA, Hilton, Accor, Marriott, IHG, Compass Group, Iberostar, and Mandarin Oriental trust us to cut their food waste by over 50% on average — and when our technology is fully adopted, it reduces pre-consumer waste by 50-70% and delivers food cost savings of 2-8%. We're building something that matters, and the people who thrive here feel that every day. We're proud to have been named one of The Sunday Times Best Places to Work and ranked #4 in the FoodTech 500 (2025) — recognition of both our innovation and our commitment to building a purpose-led company where people can make a real difference. Our values Our values aren't just words on a page. They sit at the centre of all that we do and shape our behaviours and the way we show up. If they resonate with you, we'd love to hear from you and for you to help us on our mission. People of action — done beats perfect. We move fast, learn faster Equal parts head and heart — we balance urgency with long-term thinking Hungry and humble — we're ambitious about our mission, not our egos Bravely honest — we say what we think, respectfully, with colleagues and customers alike Bound by food — it's what unites us, and it's what we're fighting for People and planet positive — we're leaving both better than we found them The team You'd be joining a small, senior platform team. We work closely with the engineers who rely on our platform daily, so communication and empathy for their day-to-day pain points matter as much as technical depth. This is a team that's comfortable with ambiguity, a lot of what we're building (especially around agentic engineering) doesn't have a playbook yet, so we value people who can define the problem as well as solve it. About the role Most platform jobs mean AWS or GCP accounts and a Kubernetes cluster. This one also means a fleet of Linux IoT devices we design and manufacture ourselves, sitting in commercial kitchens around the world and streaming data back to us. There's an on-prem GPU cluster running ML workloads, and a data estate covering Aurora Postgres, MySQL, Redshift, Timescale, Mongo and DynamoDB. We use that stack to fight food waste. Our devices watch what gets thrown away in professional kitchens and tell chefs how to stop it. It works, and we're deployed in kitchens in 94 countries and counting. Your customers are our own engineers and data scientists. Your job is to make their route to production short, safe and boring. The problem we most want help with We're working out how agentic engineering actually works at Winnow: how AI agents fit into the way we build, review, test and ship, and what platform they need underneath them to be useful rather than a liability. This work has just begun. We need to design the guardrails, the golden paths, the evaluation, the permissions model and the cost controls. In large part it's a platform problem, and nobody seems to have properly solved it yet, here or anywhere else. Key responsibilities: Design and build the platform for agentic engineering at Winnow: guardrails, golden paths, evals and cost controls, so our engineers can hand real work to an agent without it going badly. Build and run the golden paths for humans too, the paved routes our engineers use to ship to production without having to become infrastructure experts. Own our AWS and GCP infrastructure as code (Terraform with Terragrunt), across ECS for services and EKS for platform components like ELK, Grafana and Airflow. DevOps: Work with development teams to improve how we deploy. Faster feedback, fewer manual steps, safer rollbacks. We measure this and we expect it to get better. Support our data science and AI workloads: Airflow pipelines, SageMaker, and GPU scheduling on the on-prem cluster. Debug the hard things across clouds, databases, containers, network and edge. Requirements Education and experience You are likely to be degree-educated having completed studies in Computer science, computer engineering, information technology, or a related field. Solid AWS experience, plus Linux and networking fundamentals you can reason from Essential skills: Infrastructure as code, Terraform or OpenTofu. Containers and orchestration: ECS, Kubernetes, or both. Bash, plus Python or Go at a level where you'd happily own a service. CI/CD pipelines you've built and maintained, not just used. You use AI coding assistants and agents day to day, and you have views on where they help, where they don't, and what would have to be true before you'd trust one with something that matters. Nice to have: Observability (OpenTelemetry, Grafana, ELK, New Relic) GitOps (ArgoCD or Flux) Airflow, SageMaker or MLflow LLM infrastructure such as inference serving, evals and LLM/ MCP gateways SecOps, container scanning (Checkov, Trivy) & secrets management Cloud and GPU cost optimisation Postgres, MySQL, Redshift beyond the basics IoT or edge fleet operations Ansible or similar Jenkins, Github Actions Nobody has all of these. Apply anyway. We'd rather hire someone curious who learns fast than someone who ticks nine boxes and stops there. Our Technology Technology is at the forefront of what we do, and the success of our company is based on our world-class technology and on finding solutions to real-world problems that have not been solved to date. Our current stack includes: Designing and manufacturing IoT ‘smart' edg