Forward Deployed Engineer

Write and review production-grade Python with the engineering discipline the job requires: typing, async, testing, packaging, Git workflow, and PR hygiene.

Work fluently in an agentic development workflow, and exercise the judgment to know when to let the agent drive and when to take the wheel.

Select models by use case and write production prompts, with a clear view of cost, latency, and when not to use an LLM at all.

Build, debug, and evaluate retrieval-augmented generation systems, and diagnose failures that come from the data rather than the retrieval logic.

Build agents that use tools, plan, and self-correct, and build MCP servers that wrap client systems rather than only consuming them.

Design and implement multi-agent systems with proper state, memory, observability, and integration into enterprise APIs, webhooks, and queues.

Deploy to cloud and on-prem environments, and explain why BFS and Insurance clients demand on-prem and how that changes the architecture.

Design and operate agentic systems that survive production, audit, and adversaries: SLOs, secrets management, prompt injection defence, chaos exercises, and runbooks.

Evaluate and harden systems with eval suites, guardrails, and red-teaming, and evidence readiness against a published production checklist.

Speak credibly with client subject-matter experts on BFS and Insurance workflows, and name where agentic AI breaks against domain and regulatory reality.

Operate as a productive Associate FDE inside a paired pod, run a discovery session, demo to a client, and document an engagement to standard.

Choose a stack for a real problem, build it end to end, and defend the choice to a senior review panel.