AI Agent Engineer

About Calvis

• Calvis is making the trillion-dollar physical security industry AI-native. Today the industry is outdated, fragmented, and run on pen and paper. With Calvis, Humans + AI work together to secure the world.
• Our marketplace connects businesses with vetted security guards who work every shift using our AI copilot. It briefs them, takes their reports, and helps them do a great job. Meanwhile, every shift generates data that continuously makes our AI better.
• We enhance guards with robots and AI video monitoring, making Calvis the only all-in-one security offering on the market.
• Clients across the country, from critical infrastructure to manufacturing facilities, trust Calvis to help keep them safe.

Why join?

• Work directly with our technical co-founder, on a tiny team (2 engineers)
• We're heavy on AI leverage: our engineers run agents concurrently, with testing and instructions baked in, and ship through Claude Code end to end.
• Complete ownership and autonomy. You own what you work on, end to end, and you decide how to build it. We have plans, but autonomy is the point: it means looking around corners and building for what's coming before it's on our roadmap.
• Everything you build lands with real security guards, in the real world, fast.
• Austin, in-person, 5x/week — sales, ops, and engineering in the same room so nothing gets lost in translation.

The role

• Build agents that understand the physical world: directing guards, robots, and sensors working together on a live site.
• Work inside our stack: Python/Django on the backend, React for the web dashboard, Swift/Kotlin on mobile, deployed across AWS, GCP, and Heroku, calling into OpenAI, Anthropic, and Bedrock.
• Run agents at scale — think dozens running concurrently — with proper testing and instructions baked into each one.
• Take what’s working and scale it. Our products have real usage and real revenue but are still early. You’ll harden them, iterate against live data, and turn v1s into systems thousands of sites depend on.

Who you are

• Worked at a high-growth startup before. You thrive in chaos, change, and uncertainty.
• You work hard, and you’re wired in. The business, the problems, and the thing you’re building stay on your mind because you actually care.
• You live in Claude Code (or similar), not just the chat app — you think in agents and workflows, not one-off prompts.
• You can tell the difference between "write deterministic code for this" and "let the model reason toward this outcome," and you know when to reach for each.
• Product mindset. You don’t just ship code, you dig into what users are actually struggling with and care about the experience you build.