Agent Operations Engineer

About Sidekick Machines

Sidekick Machines is an AI startup that builds and runs AI agents for enterprises. Most agent projects stall on infrastructure: identity, security, integrations, approvals, monitoring and cost. We built our own platform, Sidekick, on Microsoft's agent stack to handle that infrastructure. Our engineers spend their time on the work that matters, and we stay on to run every agent we ship.

Agentic AI is new, and nobody has a decade of it. We hire people with strong fundamentals and the drive to figure things out, and we teach the rest.

The role

We don't build an agent and walk away. We stay and run it, and that's one of the most important things we offer our clients. You'll keep their agents healthy in production: watching quality, controlling cost and fixing what breaks. The Sidekick platform gives you the tracing, evals and controls to do this across many agents at once, so you spend your time improving agents, not building tooling.

If you've worked in managed services or application support at a consultancy, you already know most of this job. The difference is that the systems you run are AI agents.

What you'll do

• Track evals for the agents you look after, and catch quality drift before users notice.
• Engineer cost: find where agents spend too much and bring it down without hurting quality.
• Handle break-fix: trace a failure to its cause (a prompt, a tool, an integration, a model change) and fix it.
• Test and roll out model, prompt and tool updates safely.
• Report to clients on how their agents are doing, in plain business terms.

What we're looking for

• 3-5 years of experience in managed services, application support, SRE or DevOps, ideally at a consultancy or systems integrator.
• Strong troubleshooting. You're comfortable digging through logs, traces and metrics.
• Enough scripting (Python or similar) to automate checks and ship fixes.
• Comfort with numbers: usage, cost, SLAs.
• Calm, clear communication with clients, especially when something's broken.

Bonus

• Exposure to LLM apps, evals or observability tools.
• Experience with cloud cost optimization.

Training

We'll teach you how agents fail and how to fix them, and you'll shadow the team on live client agents before you look after your own.

Compensation and how to apply

• Compensation: Salary plus ESOPs
• To apply: [HIDDEN TEXT]