Senior AI Engineer, Agentic Platform

New

WorkWaveAI analytics

Remote (UK-based)Full-TimeSenior

Salary not disclosed

Apply NowOpens the employer's application page

Job Details

Experience: 5+ years building production software or data-intensive systems used by real customers
Required Skills: AWSPythonSQL

Have 5+ years of experience building production software or data-intensive systems used by real customers.
Have strong Python skills and have shipped at least one LLM or agent feature to production users.
Have hands-on AWS experience; Amazon Bedrock and Bedrock AgentCore experience is preferred.
Have designed agent orchestration, tools and functions, context assembly, and multi-step workflows.
Have built retrieval over real business data and can explain context selection and why an answer was wrong.
Have enough SQL and metrics fluency to identify when an agent uses the wrong number.
Have built or owned an evaluation loop for a generative system, including labeled sets, regression gates, and defensible measures of improvement.
Have product judgment about when an agent should answer, ask, or refuse, and how it should show its work.
Have quantified agent outcomes such as adoption, task success, or decision quality, and presented findings to non-technical audiences.
Have used coding agents on production work and can verify correct output and build harnesses for future tasks.
Have excellent communication and collaboration skills and experience working across technical and non-technical teams.

Build production agents and the tools they call, including retrieval, multi-step workflows, and action-taking, using Amazon Bedrock and Bedrock AgentCore.
Turn semantic layers and service APIs into endpoints and tools agents can use, selecting the metrics and entities they should access.
Shape agent experiences, including answer speed, confidence, explanations, and suggested next steps.
Build evaluation harnesses with labeled test sets, groundedness and hallucination measures, and CI regression gates.
Measure adoption, feedback, intent, and realized customer impact, and report results to Product and the business.
Instrument tracing, diagnose answer failures, and turn failure modes into fixes.
Own the serving path, including MCP endpoints, latency, cost per response, caching, runtime guardrails, deployment, and testing.
Partner with Product Managers and engineers to define which decisions agents should support and drive data or infrastructure changes with the owning team.

View Full Description & ApplyYou'll be redirected to the employer's site