- Work with clients to understand business problems and translate them into scalable, production-ready agentic AI solutions.
- Design and implement multi-step agent systems, including orchestration, tool and function calling, task decomposition, and multi-agent coordination where appropriate.
- Integrate agents with enterprise systems by exposing APIs as tools, implementing MCP servers, and developing structured tool contracts.
- Build evaluation frameworks, guardrails, grounding and citation checks, and human-in-the-loop workflows.
- Instrument agent systems for traceability and observability, including actions, model behavior, token and cost usage, latency, and performance.
- Design and maintain state and memory for long-running workflows, and implement model routing, retries, and idempotency.
- Build and deploy AI solutions using Azure or AWS cloud platforms and services.
- Apply software engineering practices using Python or Java, APIs, containers, source control, automated testing, and CI/CD.
- Work with client engineering and security teams on architecture reviews, threat modeling, security reviews, audits, and enterprise delivery requirements.
- Assess when agentic architectures are appropriate versus deterministic pipelines, traditional machine learning, or other approaches.
AWSPythonJava+2 more