Job Description:Overview:We are seeking an AI Agent Developer to design, build, and deploy AI agents that transform how teams at Chatham work. This is a hands-on engineering role for someone who wants to apply modern AI to real business problems and see the impact firsthand. You will report to one of our Software Engineering Directors to maintain alignment with our organizational AI objectives.
In each role, you will embed with a partner team as a forward-deployed engineer. You will review past work, shadow team members on live tasks, and map where time is spent and why. From there, you will identify and prioritize the highest-value opportunities, then build, evaluate, and release agents that triage requests, research and draft answers, investigate issues across systems and data, and prepare work for review. The team's experts stay in the loop throughout, and people review agent output before it reaches a client or a production system.
Once a team's agents are running reliably, you will hand them over to the people who own the work and bring the same approach to the next team. Over time, you will partner with teams across the firm, from engineering and operations to client service and Marketing, wherever repetitive work can be improved through agentic means. You will work closely with Chatham's AI evaluation and platform teams and help set the standards that other teams adopt.
This is a new role at Chatham and an opportunity to help define how the firm builds and scales agentic capabilities. The ideal candidate is a pragmatic engineer who enjoys learning new domains from the people who do the work, collaborates easily with technical and non-technical colleagues, and measures success by the problems solved for users.
In this role you will:Discovery & Partnership
- Embed in a partner team and learn how it works. Review past work, shadow live tasks, and map the steps that the team repeats.
- Find and rank the tasks where agents can save the most time or improve quality and agree on priorities with the team's leaders.
- Work closely with the team's experts. They supply the domain knowledge and test cases, and they review agent output.
Agent Development
- Design, build, and deploy cloud-native AI agents that do repetitive work from documented processes, with a person in the loop where judgment is necessary.
- Build agents that triage and route incoming requests, research questions and draft answers, and investigate issues across systems, logs, and data.
- Build agents that prepare work for people to review, such as draft responses, reports, and code changes as merge requests.
- Turn each team's playbooks and procedures into agent instructions and tools and keep them current as processes change.
Evaluation & Quality
- Create evaluation sets from real past work. Measure each agent's accuracy before release and whenever models or techniques change and monitor the results in production.
- Partner with our AI evaluation team, and adopt its tools, techniques, and processes as they mature.
- Follow Chatham's standards for security, data handling, and the responsible use of AI.
Reuse & Handover
- Document each agent's purpose, tools, and limits so that other teams can adopt and extend it.
- Build reusable tools and patterns that make each new team's agents faster to deliver.
- Hand each team's agents over to the people who own the work, then move to the next team.
Your Impact:For each partner team
- Agents do a growing share of the team's routine work correctly, and people review the output.
- The team spends less time on repetitive tasks and more time on work that needs judgment.
- Requests that wait in a queue today get faster and more consistent answers.
- Each agent meets its accuracy targets before release and keeps them in production.
Across Chatham
- Each agent has documentation and tests, and other teams reuse its tools and patterns.
- Each engagement ends with a clean handover to the people who own the work.
- A ranked list of agent opportunities is ready for the next team before you move to it.
Contributors to your success:- 3+ years of professional software engineering experience.
- Strong proficiency in Python or C#, with experience in APIs, services, or automation.
- Hands-on experience with applications built on large language models (LLMs), including prompt and context design, tool use, and agent workflows.
- Experience with evaluation of AI output: test sets, accuracy measures, and regression checks for systems that do not always return the same answer.
- Solid SQL skills, and comfort with logs and telemetry to find the root cause of a problem.
- Experience with Git, code review, and CI/CD.
- Curiosity and patience to learn complex business domains from the people who do the work.
- Ability to turn a business process into clear steps that an agent can follow, and to judge where a person must stay in the loop.
- Clear written and verbal communication with technical and non-technical colleagues.
- Experience with tools and integrations built on the Model Context Protocol (MCP).
- Experience with platforms or frameworks for AI agents, such as Microsoft Foundry or the Claude Agent SDK.
- Daily use of AI coding tools such as Claude Code, Codex or GitHub Copilot.
- Experience in a forward-deployed, solutions engineering, or internal tools role, or a background in operations, technical support, or site reliability engineering.
- Familiarity with tools such as GitHub, Grafana, Jira, and Confluence.
