AI​/Agentic Engineer

We are seeking AI / Agentic Engineers to design and build production-oriented AI agents that address client pain points and improve service experiences. The team has already analyzed service requests, identified recurring themes and intents, and prioritized high-value agentic opportunities. Several agents are approaching launch, and the broader roadmap includes many largely greenfield solutions with limited reuse.

This role is for a strong software engineer who can reason through ambiguous, nondeterministic problems—not someone whose AI experience is limited to prompt engineering or using copilots. The engineer will independently translate business problems into clean abstractions, API contracts, data models, tool interactions, and observable agent workflows that can be tested, improved, and trusted over time.

Design and build greenfield AI agents that support client workflows or serve as tier-zero responders to incoming service requests.

Translate loosely defined business problems into well-structured agent behaviors, system boundaries, abstraction layers, and contract-driven interfaces.

Define APIs and communication contracts among agents, tools, services, and supporting components.

Develop Python-based services and agent workflows with strong data modeling and validation practices.

Integrate agents with enterprise APIs, tools, MCP-based interfaces, model APIs, memory systems, retrieval components, and cloud services.

Design for production reliability across retries, timeouts, failure handling, idempotency, back pressure, state, and distributed communication.

Implement observability and evaluation approaches that measure agent correctness, trace behavior, detect regressions, and verify that prompt, model, tool, or workflow changes do not degrade outcomes.

Iterate rapidly and deliver working increments within sprints while requirements and solution patterns continue to evolve.

Collaborate with engineering and business partners, explain design decisions clearly, and occasionally demonstrate agent capabilities or answer technical questions in client discussions.

Strong Python software-engineering experience.

Demonstrated strength in API design, abstraction, contract-driven development, and defining clean boundaries between system components.

Strong data-modeling and validation skills; experience representing business problems in clear, enforceable schemas.

Practical understanding of how LLM-powered agents operate, including model interaction, tool selection and execution, the tool loop, context, state, and memory.

Ability to reason about nondeterministic behavior and design solutions that can be evaluated, observed, and improved rather than merely confirming that an agent produces a response.

Experience building or integrating backend services and distributed systems using patterns such as REST, RPC, Web Sockets, or gRPC.

Understanding of production concerns such as retries, timeouts, error handling, back pressure, and service reliability.

Comfort working independently through ambiguity, asking the right questions, and making sound technical decisions without line-by-line direction.

Ability to work in a fast-moving environment with continuous, sprint-level deliverables.

Strong communication skills and sufficient professional polish to explain or demonstrate a solution to internal stakeholders and, when needed, clients.

Hands-on experience building production AI agents or multi-agent systems.

Experience with AWS AI and application services, particularly Amazon Bedrock, Agent Core, EKS, SQS, and DynamoDB.

Experience with Anthropic or OpenAI model APIs.

Experience with Pydantic for Python data modeling and validation.

Familiarity with Lang Chain, Lang Graph, Amazon Strands, or comparable agent frameworks.

Experience with agentic design patterns, including head-agent/sub-agent orchestration and agent-to-agent communication.

Experience with retrieval-augmented generation, context or retrieval graphs, and agent memory architectures.

Experience designing agent evaluation, tracing, monitoring, observability, and regression-testing strategies.

Understanding of MCP and other mechanisms for exposing tools to agents.