Agentic AI Engineer

Agentic AI Engineer - build an AI agent framework from zero There are no autonomous agents running in production here yet. That's the job: build the first ones, and define how every agent after them gets designed, governed and shipped. You'd be the founding engineer on a brand-new Agentic Workforce team, reporting into the VP of Data \& AI at a fast-growing global SaaS business in the risk, compliance and safety-management space. The business has scaled rapidly through acquisition, growing from a few hundred people to a couple of thousand in a few years, and is now investing seriously in agentic AI as a core part of its technology roadmap. The problem you'd be solving Right now, agent-building across the business happens in three tiers: individual agents people build for their own workflows, team-level agents built for shared use, and "gold standard" agents that operate autonomously in high-risk, regulated parts of the business, think compliance and safety documentation. That third tier doesn't exist yet. You'd build it: designing, deploying and governing production-grade agents that orchestrate LLMs, tools, memory and business logic, while also setting the evaluation, guardrail and observability standards the rest of the business builds against. It's not a purely heads-down build role. You'll work directly with teams across the business to turn real operational pain points into shipped agentic solutions, so the ability to translate a messy, half-formed business problem into a working system matters as much as the engineering itself. What you'd bring * Strong Python and backend/API foundations * Hands-on experience building LLM-based agentic systems with orchestration frameworks like LangGraph, LangChain, Microsoft Agent Framework or CrewAI * Comfortable designing agents that reason, plan, use tools and maintain state across multi-step tasks * Experience with prompt engineering, embeddings and RAG * Some grounding in agent evaluation, guardrails or observability, since safe production behaviour is a first-class part of this role * Cloud deployment experience (AWS preferred, not essential) and familiarity with containerisation (Docker, Kubernetes or similar) * Comfortable working directly with non-technical stakeholders to shape requirements, not just execute a spec Hybrid across a number of locations, or remote-first with monthly travel.