Forward Deployed Engineer Intern
Yes, you will meet the client. Work with the people doing the job, find the manual work and build the agent that takes it away. Your first pull request merges in week one.
This is not the internship where you fix the printer.
Forward deployed means you work close to the problem: with the client's team, on their real workflows and data. Sometimes that is in their office. More often it is a call, a screen share and a shared channel. Then you build the fix and stay with it until it works.
What we are building
TenthNode builds AI agents that take manual work off operations teams: the tickets, the spreadsheets, the copy-paste between systems and the "just following up" emails.
We started in property and hospitality, where the founders spent their careers. The same work shows up in every business that runs on people and systems. We are early, so whatever you build or write gets used.
What you will actually do
- Work with the people who do the job today. Learn how the work really gets done, not how the process map says it does.
- Write a short brief on what hurts, what it costs and what we would build.
- Build agents that read, decide and act inside real business workflows: invoices, enquiries, applications and reports.
- Ship small pull requests every day. A founder reviews each one before it reaches a client.
- Decide where the agent stops and asks a human. That line matters more than the model.
- Stay with it after launch. Watch how people use it and fix what slows them down.
- Sit in on client conversations with the founders. They have spent nearly 40 years between them watching business systems fail. Learn from their scars instead of collecting your own.
What you will learn
- How to build with AI agents properly, not prompt and pray.
- How to get the real problem out of an operator who is short on time and has heard every pitch.
- How real operations software works and why it is harder than it looks.
- How an early company decides what to build and what to kill.
- Zero coffee-run skills. Sorry.
Who you are
- You have built something real and the commit history proves it.
- You can explain every line of your own code, including the lines an AI wrote.
- You can explain what you built to someone who does not code, such as an operations lead or a business owner.
- You ask questions before you start coding.
- You already live in AI tools and have opinions about them.
- You ship first, polish second and say so when something is broken.
- A degree is nice. Proof of building is nicer.
What you will build with
We pick tools for the job. Tell us when a better one shows up. Today that means:
- Python or TypeScript
- Postgres, such as Supabase
- Workflow tools like n8n
- LLM APIs from OpenAI, Anthropic or Google, plus tool calling, MCP and evals
- Git and pull requests
- Cloud deploys on Railway, AWS or Azure
- Microsoft 365 and Graph, since many clients run on Microsoft
You do not need all of it. You need one language you know well and the habit of picking up the rest.
How we hire
- We read your commits.
- A 30-minute call where you walk us through your own code.
- A short assignment on sample data.
- A conversation where you work through a real problem with us.
You already build with AI coding agents, you have shipped something people outside your class use and you have caught yourself saying "I wonder if an agent could do that."
Find the busywork.Build the agent that ends it.
Brand & Content Marketing Intern
Make AI agents at work make sense to people who have never opened a prompt box. Your name on published work.
Venture Scout
Introduce a business that still does work manually. If it signs and pays, you earn 20% of the first engagement fee. Commission only.