Gauntlet AI Fellowship

What Gauntlet AI is

Gauntlet AI is a ten-week, all-expenses-paid immersive program for the most talented engineers in the world. You do deep project work with a group of peers who are just as sharp as you, pushing the limits of what you can build with AI.

Gauntlet AI covers your housing, flights, food, laundry, and compute so nothing gets in the way of becoming a great AI engineer. At the end, we help you land an AI engineering role with one of our hiring partners. Roles start around $200K, and more experienced engineers can earn much more.

The program is completely free for you. Forever. No tuition. No bill later. No cut of your pay. Companies pay us a recruiting fee to hire engineers who complete the program.

How Gauntlet AI started

Gauntlet AI has trained software engineers since 2017. We started by taking people from zero experience to their first software engineering job.

In 2023, we sent a research team to some of our largest hiring partners to understand how AI was changing software development. One of those partners was Trilogy, which operates 160 software companies and employs about 1,200 engineers worldwide.

That research led us to create our first AI training program, AI for Developer Productivity. The eight-week virtual program trained engineers across 70 companies to build AI agents that automate key parts of the software development process, including code review, QA, architecture design, and technical writing.

Trilogy alone sent 100 engineers through the program. Afterward, they saw a 50% quarter-over-quarter increase in productivity among the engineers who participated.

In 2024, we built on that success and created Gauntlet AI. We made it onsite, intensive, and all-in. Our hiring partners needed AI-first talent, so we built a program that trains engineers, matches them with companies, and helps them land AI engineering roles.

Since then, we’ve trained 1,000+ software engineers across 100+ companies and government agencies. Our alumni code at Trilogy, Carvana, Nerdy, Function Health, TerraFirma, Alpha School, the U.S. Treasury, and many more.

How Gauntlet AI works

Gauntlet AI is a 10 week intensive program: 3 weeks remote and 7 weeks onsite in Austin, Texas.

The program has three stages.

Prove you can build

WEEKS 1–3 | REMOTE

The first three weeks are remote and focused on vetting.

We only want the strongest engineers to make it onsite to Austin.

You learn to build with coding agents and orchestrate teams of agents that build for you. You work with RAG, graphs, agents, and evals. Every day, you practice building a software factory, a system where you manage multiple AI agents working together to build software.

We also complete checks that our hiring partners would normally do during the first or second round of their hiring process, including a background check and a proctored CCAT.

By the end of Week 3, you’ve proven you can build at the speed and level we expect, and we’ve confirmed your experience.

Get hired

WEEKS 4–6 | ONSITE

The next three weeks are focused on placement.

Our hiring partners give you projects that simulate the work you would do if they hired you. You might build a system, solve a technical problem, or work through a challenge based on a real problem at their company.

You complete projects for the companies you’re matched with. Those companies come onsite for Hiring Partner Days, where you present what you’ve built and interview with their teams.

On average, challengers complete 7.2 interviews during this part of the program.

Push the frontier

WEEKS 7–10 | ONSITE

The final four weeks go deeper into how AI itself is built.

You train a language model from scratch. You work with small language models, PEFT, and LoRA. You benchmark models and learn how to decide which model is best for a given job.

You also go beyond software development. You build marketing factories, create eval pipelines, and learn AI development processes like spec-driven development.

Everything leads to your capstone, a contribution to the open-source AI community. It could be a model, agent, pipeline, tool, or something we haven’t seen before.

By the end, you know how to build, evaluate, and ship AI systems that work in the real world.

We call everyone in the program a challenger. That’s because every week, you have to complete a technical challenge to stay in the program. These challenges simulate the kind of work you’ll be expected to do as an AI engineer.

If you can’t complete a challenge, you leave the program. Gauntlet is completely free, so your skin in the game is your seat. As long as you perform, you stay.

How to get into Gauntlet AI

To join our next cohort, you need at least 3 years of software development experience. This can include full-stack, frontend, backend, mobile, DevOps, data engineering, machine learning engineering, or technical leadership.

We’re looking for engineers who have experience building and shipping production software. Gauntlet AI is not designed for product managers, data analysts, founders without a technical background, or people looking to move into AI from a non-technical role.

Every applicant also takes the CCAT, a 15-minute problem-solving assessment. You need to score at least 40 out of 50 to be admitted.

The application process is simple. Fill out the application below and we’ll review your LinkedIn for the required experience. If you meet our bar, you’ll move to the second round and take the CCAT.