DevOps Engineer

Location & work mode

This role is based in Bengaluru, India, and works from our office. We care about focused time and short feedback loops, which is easier in a room together while the team is this size.

What you'll do

• Run the GPU inference fleet the API serves from: capacity, autoscaling, rollout, and what happens when a node dies mid-request.
• Keep time to first audio low in production, not just in a benchmark — that means real work on cold starts, model loading, and request routing.
• Own deployments end to end: containers, CI/CD, and releases that do not need a person watching them.
• Build the observability that makes an incident short — metrics, logs, traces, and alerts that fire on things that matter.
• Handle multi-region: we serve from the EU, the US and India, and Enterprise customers ask about residency.
• Keep the platform secure and the secrets where they belong.

What we're looking for

• Experience running production infrastructure that people depend on, including the on-call side of it.
• Real Kubernetes and container experience, and a clear sense of when not to reach for them.
• Comfort with GPU workloads: scheduling, drivers, memory, and why the second request is faster than the first.
• Infrastructure as code, and CI/CD you have built rather than inherited.
• Enough Python to read and debug the services you are deploying.
• Clear writing. We work from written arguments more than meetings.

Bonus

• You have served ML models at scale — vLLM, Triton, or something you had to write yourself.
• You have worked with streaming or low-latency systems, WebSockets included.
• You have run infrastructure across more than one cloud or GPU provider.
• You have been the person who made a slow deploy fast.

What we offer

• Competitive salary, discussed early in the process.
• Direct ownership of a system in production, not a slice of a backlog.
• A small team and a short path from idea to shipped.
• Hardware and compute you need to do the work.
• Based in our Bengaluru office.