Forward Deployed Engineer Intern

What you will do

Map a business workflow before you build. For example, how a company onboards a new customer: the steps, the systems involved, the approvals, the exceptions and the handoffs between people.

Connect AI agents to the systems a business already runs, such as a CRM, a data platform, a document store, email, WhatsApp and internal APIs, so the agent reads and updates the right data.

Design the orchestration: break a workflow into steps and decide which steps an agent handles, which call another system, and which go to a person for approval.

Build multi-step and multi-agent workflows that pass work between agents, tools and people, and recover when a step fails.

Build retrieval over company documents and data, so agents answer from the client's own information.

Give agents access to tools through APIs and the Model Context Protocol (MCP).

Test agent behaviour with evaluation sets that catch wrong answers, missed steps and unsafe actions, and run them before every release.

Track the cost and response time of each workflow, and reduce them through prompt design, caching and choosing the right model for each step.

Join client working sessions, demo what you built, and help take it to go-live.

Your first 90 days

Month 1: Build a small agent workflow that connects two systems on our internal stack. Learn how we test agents and track cost. Sit in on client working sessions.

Month 2: Build one part of a client project, such as a system connection, a workflow step or an evaluation set, with review from a senior engineer.

Month 3: Take that part from requirement to go-live with the team.

What you need

Strong coding in Python or TypeScript, and comfort reading code you did not write.

At least one project built on a large language model API that uses tool calling, retrieval or an agent framework. Send us the link.

Experience with REST APIs, webhooks and JSON, and with making one system talk to another.

A working understanding of tokens, context windows, rate limits and latency.

Basic SQL.

Clear written and spoken English. You will talk to clients directly.

Availability to work full-time for 6 months.

Good to have

Experience with Salesforce or Databricks.

Experience with an agent framework such as LangGraph, the OpenAI Agents SDK or the Claude Agent SDK.

You have built or used an MCP server.

Experience with workflow automation tools or event-driven systems.

Experience deploying a service on AWS, Azure or Google Cloud, including Docker.

Daily use of AI coding tools such as Claude Code or Cursor.

What you get

A paid internship.

Direct mentoring from the founders and senior engineers.

A full-time offer as a Forward Deployed Engineer after the internship, based on your performance and evaluation.

How to apply

• Apply on LinkedIn with your CV and a link to one thing you built: a GitHub repository, a demo video or a deployed app.
• A 30-minute call with our team.
• A take-home task of about 3 hours. You will get a short description of a business workflow, build a small agent that connects two systems to complete it, and report what one run costs in tokens.
• A 60-minute interview with the founders. You will walk us through your task, and we will role-play a client conversation.
• A decision within 10 working days of your application.

Internship details

Start date: October 2026