Senior AI Engineer

What You'll Do

Design the system, not just the prompts

• Architect the multi-agent system — design how our agents work together coherently across the product.
• Own reliability of AI workflows — tackle race conditions and edge cases across a large, concurrent surface area.
• Manage inference cost — optimise model usage and architecture to keep costs sustainable as we scale.
• Build evaluation & guardrails — make agent behaviour measurable, testable and safe.
• Partner with full-stack & platform — ship AI features that are robust in production, not just in demos.

Who You Are

An AI engineer who ships to production

• Strong applied AI / ML experience — you've built and shipped LLM-based or agentic systems in production.
• Systems thinking — you architect holistically and reason about concurrency, race conditions and cost.
• Multi-agent / LLM orchestration — hands-on with agent frameworks, tool use, and evaluation.
• Solid software engineering — comfortable in Python (and ideally Go/TypeScript) writing production-grade code.
• Pragmatic & startup-minded — you balance ambition with what ships reliably this quarter.

About the Role

Architect the AI at the core of the product

Paul's Job is an AI-native recruiting platform replacing fragmented legacy software with one unified system that automates candidate communication, screening and interview scheduling. We have real traction — customers in rollout and a clear Series A path — and a lean six-person engineering team shipping fast across Go, Python and TypeScript.

AI is the core of Paul's Job — including Smith, our AI implementation agent, and the agents that automate candidate communication, screening and scheduling. The surface area is massive and growing, and we need someone to architect the multi-agent system holistically rather than bolting on intelligence service by service, while tackling race conditions and cost.

Why Paul's Job

Own the AI that defines us

• Join a small, high-trust engineering team where you own real surface area from day one.
• Work on an AI-native product heading into Series A with genuine traction.
• Architect the agentic core of the product, not a peripheral feature.
• Remote-friendly, fast-moving, and free of legacy bureaucracy.