Forward Deployed Engineer

About this role

About the Role

Clutch builds AI agents that work on behalf of credit unions. Emma, our collections agent, is live at five credit unions and handled 60,000+ member conversations in her first twelve weeks at one of them. Behind her is a growing portfolio: origination, fraud step-up, member (re)engagement and activation agents, all built on the same Agents Platform (TypeScript, NestJS, our own agent SDK, MCP, Temporal, promptfoo).

As a Forward Deployed Engineer you are one of the two engineers on an agent. You take it from "sold" to "live and performing" at a specific credit union: you sit in the discovery sessions, you build the integrations to their core banking and loan systems, you author and tune the agent's behavior, you run the live tests with their staff, and you stay with it through go-live and the first weeks of production. When something needs a person in the room, you are the one who flies. You are based in the US and can be at a credit union in the Midwest or the Southeast within a day or two, not a week.

This is a senior engineering role, not a solutions or support role. You write production code in our monorepo every week: integrations, modules, and the agents themselves (prompts, tools, evals), together with the other engineers on the pod. The difference from a traditional engineer is that you are on site more often, you make judgment calls in front of the customer, and you own the outcome. What you learn in the field flows straight back into the product: defaults, playbooks, and the platform capabilities the next agent will need.

You build; our AI Operations Managers deliver. The pod (pod lead plus engineers) designs and builds each agent in a high-touch way with the first customers. Once the agent is defined and live, AI Ops coordinates the implementations, configuration and ongoing success at the next credit unions. You hand them something that works and a playbook to run it.

About the Team

You'll join the Agents team, a small, senior group operating as a startup inside Clutch: seven engineers, a dedicated AI product group, and AI sales specialists, working directly with some of the largest credit unions in the country. Every agent is staffed by a pair of engineers, one leaning toward the conversation and LLM side, one toward software and integrations, with a pod lead who owns the product day to day. There is no layer between you and the customer, and no layer between you and the platform team.

The operating style is direct, written, and evidence-driven: BLUF updates, evals over opinions, business outcomes over ticket counts. The team is spread across Brazil, Argentina and the US and works async by default. Being on site is a tool we use at specific moments (kickoff, live testing, go-live), not a permanent posting.

What You’ll Do

Within 1 months, you will:

- Ship your first production change to a live agent through the team's review process, and be able to explain how a member's call flows through the Agents Platform end to end

- Learn one agent deeply: system prompt, tool contracts, orchestration triggers, compliance gates, and the implementation process from kickoff to go-live

- Join at least one credit union implementation as the second engineer, including a live testing session with their staff

- Map where the current implementation process loses time and where a default or a tool would have removed a round trip

Within 3 months, you will:

- Own the technical side of one credit union implementation end to end: core and LOS integration, agent configuration, testing, go-live, and the first weeks of production

- Run the eval and tuning loop for your agent (golden transcripts, rubric judges, regression suite) and be trusted to change its behavior without a second engineer checking your work

- Have completed at least two on-site visits and turned what you heard there into product decisions with your pod lead and PM

- Contribute at least one reusable capability back to the Agents Platform that the next implementation will not have to rebuild

Within 6 months, you will:

- Have taken one net-new agent from sold to live at its first credit union, with data you can show the partner

- Be the engineer credit union staff ask for by name when something in their deployment needs to change

- Have written the implementation playbook for your agent well enough that the AI Ops team runs the second and third credit union without you

- Be shaping what the next agents look like, because you have seen where the current ones break in the field

What You’ll Bring

In rough priority order. Prior LLM or ML experience is not required. We can ramp the agent skills; we cannot ramp senior engineering judgment.

- 5+ years of software engineering in a fast-paced environment, preferably in SaaS and/or Fintech, with senior TypeScript and Node.js backend experience in a large codebase (NestJS, CQRS, schema-validated boundaries, dependency injection). We do not count years of AI experience; the field is too young for that to mean anything. We look at what you have shipped and how you reason about it

- Integration and API engineering: secure service-to-service endpoints, third-party APIs, webhooks, and the patience to work with legacy banking systems that were not designed to be integrated

- Security and auth judgment: OAuth and machine-to-machine auth, tenant-scoped access, secrets handling, no PII in logs. Our agents touch sensitive member data and operate under consumer-contact regulation

- Customer presence: you can sit across from a VP of Collections, hear a request, and say "here is what I recommend and why", including when the answer is no. You are comfortable making calls in the room and owning them afterwards

- Code-quality judgment under deadline pressure: you take on debt deliberately, name it, and pay it down, rather than gold-plating or leaving silent debt behind

- Willingness to travel: roughly 20% on average, with one- to two-week…