Applied AI Engineer, Agents & Automations

Cohere Cohere · AI Frontier · EUROPE · Platform

Cohere is seeking an Applied AI Engineer to build and improve AI-powered product experiences within their North platform. This role focuses on making AI agents reliable, useful, and production-ready for enterprise customers, involving the design of interfaces, building evaluation and feedback systems, and improving AI reliability through product design and prompting. The engineer will work across backend, product, data, and model behavior, shipping AI-assisted product experiences into production.

What you'd actually do

  1. Build AI-powered product experiences that help users create, configure, evaluate, and improve workflows across North.
  2. Design and ship new interfaces for interacting with models, including workflow builders, proactive assistants, review flows, and other human-AI collaboration patterns.
  3. Build eval, observability, and feedback systems that measure whether AI experiences succeed in real enterprise workflows.
  4. Improve AI reliability through product design, prompting, context construction, tool-use strategies, instrumentation, and feedback loops.
  5. Turn production failures, user feedback, and real-task data into better product behavior, stronger eval suites, regression tests, and signals for modelling.

Skills

Required

  • Strong software engineering skills
  • Experience shipping LLM-powered or agent-powered product experiences
  • Experience improving AI products based on real-world feedback
  • Experience debugging real-world failures
  • Experience working across system boundaries

Nice to have

  • Experience with backend systems, user-facing product, data, evals, and model behavior
  • Product taste and understanding of how interface, workflow, and model interact
  • Experience with prompt and context experiments
  • Experience with tool-use strategies
  • Experience with observability and feedback loops

What the JD emphasized

  • shipped LLM-powered, agent-powered, or AI-assisted product experiences into production
  • improved them based on real-world feedback
  • debugging messy real-world failures
  • turning what you learn from production, user testing, and evals into durable improvements
  • owning ambiguous problems end-to-end

Other signals

  • building AI systems
  • deploying frontier models
  • making agents genuinely useful
  • improving AI reliability
  • shipping AI-powered product experiences