AI / Agent Engineer

The role

Key responsibilities

• Develop AI agents and workflows based on approved business requirements and technical designs.
• Integrate language models through model APIs or approved deployment environments.
• Design and maintain prompts, structured outputs, tool definitions, and workflow logic.
• Implement retrieval-augmented generation, including document processing, indexing, retrieval, and source attribution where required.
• Connect agents to approved tools and data sources in collaboration with the integration engineer.
• Implement workflow state, context management, retries, fallbacks, and exception handling.
• Establish human-review and approval steps for sensitive or consequential actions.
• Build evaluation datasets and automated evaluations for task completion, factual accuracy, output quality, and adherence to instructions.
• Test for failure modes such as unsupported claims, prompt injection, inappropriate tool use, and unintended data disclosure.
• Monitor and improve response quality, latency, token usage, and model costs.
• Maintain version control for prompts, workflow configurations, evaluation datasets, and model settings.
• Document agent behavior, limitations, dependencies, and operating requirements.

Required experience and capabilities

• Strong programming skills in Python, TypeScript, or a comparable language.
• Practical experience developing applications using large language models.
• Experience with model APIs, structured outputs, function calling, or tool-enabled workflows.
• Understanding of retrieval, embeddings, vector search, and document-processing pipelines.
• Ability to design reproducible tests and evaluate probabilistic outputs.
• Experience with APIs, databases, source control, and automated testing.
• Strong debugging skills and an analytical approach to error investigation.

Preferred experience

• Production deployment of agent-based or AI-assisted applications.
• Experience with workflow orchestration or agent-development frameworks.
• Familiarity with AI observability, evaluation tooling, and model benchmarking.
• Experience automating marketing, sales, customer support, or internal operations.