About the role
Spyne builds AI for US car dealerships: Vini AI agents that answer their phone calls, texts and leads, and Studio AI, which turns photos of cars into listing-ready images and video. When our platform is slow or down, a rooftop's listings and phones are affected during its selling day.
You own reliability, deployment, observability and infrastructure cost across all products. Product teams own their services; you own the platform, the standards, and the incident process they run on. Our production is not a plain web stack: Studio runs GPU and media-processing pipelines on thousands of photos and videos, and Vini runs real-time voice, where latency is the product. You are a builder first: you use agentic tools to automate the work others do by hand, and you set the standard for how the team runs production.
What you'll do
- Own uptime, incident response and root cause analysis for production, including vendor failover and degradation paths for third-party telephony and model APIs, and a fair, sustainable on-call model.
- Build and run the deploy pipeline, environments and rollback paths.
- Own platform security: secrets, access control, and [compliance scope].
- Own observability: logs, metrics, alerts and dashboards that catch problems before a customer or CSM does.
- Own cloud cost: a published plan, a dashboard that shows progress, and savings that stick as the rooftop count grows.
- Automate toil with agents and scripts; build runbooks and let the rest of engineering self-serve.
- Build and lead a small DevOps/SRE team over time; hire and raise the bar.
What we look for
- You have run production for a product customers pay for, and led or been on the hook for its on-call.
- Deep infrastructure knowledge: AWS, containers and orchestration, CI/CD, infrastructure as code, networking, databases in production.
- Strong agentic builder: you automate with Claude Code or similar tools and can show examples.
- You stay on an outage until the cause is proven and prevented, and you tell people early when something is going wrong.
- Proven incident leadership: calm under pressure, clear written post-mortems, fixes that remove the class of problem.
- Cost discipline: you have cut cloud spend without hurting reliability.
- Security and secrets management in production.
Nice to have
- GPU or media-processing workloads, voice or real-time systems.
Perks
- Pay: market-leading salary and stock options.
- Health and life cover: group medical (GMC), personal accident (GPA) and term life (GTLI) insurance for you and your family.
- Learning: LinkedIn Learning access and mentorship programs.




