Ten weeks. No shortcuts.
Ten weeks of sustained execution that resets how you build. You emerge able to design and ship production-grade AI systems that advance your career.
Tell me more about Gauntlet
The Problem
You're already reading the papers, experimenting between meetings, pushing yourself to stay ahead. That energy is real — but scattered effort won't close the gap. You need a forcing function to go from curious to dangerous.
Gauntlet exists because the gap between "experimenting with AI" and building with it as a core primitive doesn't close gradually. It closes in a compressed burst of focused execution.
Austin, TX — Live instruction
The Format
Gauntlet is a controlled environment for sustained execution — where every week produces visible proof of what you can build.
Travel, housing, food, compute, model access are paid for by hiring partners looking to recruit Gauntlet talent.
Full-time from day one. Establish pace, tool fluency, and architectural discipline before the real pressure starts.
Relocate. Build in-person. The proximity, intensity, and accountability change how you operate.
Companies watch you build in real time. Not your résumé — your output, your reasoning, your response to pressure.
Challengers use the latest models to extend their skills and push what is possible.
Ship every week. Real constraints. Real deadlines. No hypotheticals.
Proof, Not Promises
Companies don't guess about Gauntlet graduates. They've already watched you ship — under pressure, with rising standards, for ten straight weeks.
0
First-round interviews secured
$200k - $950k
Starting compensation at or above
0
Weeks of sustained execution
Graduates joined companies including
"I came in thinking I was strong. By week four, I realized I'd been building with one hand tied behind my back. By week ten, I couldn't imagine going back.
Cohort collaboration
Every Single Week
Monday: assignment drops. Friday: you ship. No extensions. No excuses. The standard rises every week. So do you.
By week three, you're building things that would have taken you months. By week ten, you've forgotten what "slow" felt like.
Weeks 1–10 · Escalating complexity
The Curriculum
Two phases. Escalating complexity. By the end, you're building systems that would have been unimaginable at the start.
An immersive launch into AI-first engineering where challengers learn to build, iterate, and ship at speed through intelligent workflows and guided collaboration.
An introduction to AI-first engineering where challengers establish core workflows, coding alongside intelligent agents and accelerating creation through AI collaboration.
Claude Code, Cursor, Codex, MCP integrations
Spec-driven development, understanding how MCPs accelerate development, building complex collaborative systems
Focus shifts to applying AI-first workflows to unfamiliar and enterprise-scale codebases, connecting live information with reasoning systems to modernize and extend existing software.
Vector databases (Pinecone, Weaviate, Qdrant), RAG frameworks
Selecting the right RAG architecture, adapting AI-first workflows to legacy constraints, production-oriented context design
Challengers design and ship production-ready AI agents for real-world domains, emphasizing reliability, evaluation, and system-level thinking.
Agent frameworks, evaluation and observability tooling
Systematic agent development, testing and iteration, evaluation-driven improvement
The in-person phase develops challengers into system-level AI engineers focused on delivery, reliability, and operating in real enterprise environments.
Challengers execute multiple real client projects, translating ambiguous requirements into production-ready systems under professional constraints.
Project-dependent tooling and AI-first workflows
Client collaboration, scoping, efficient execution
Focus shifts to speed, quality, and consistency as challengers deliver additional client projects using refined AI-first methodologies.
Full AI development stack and project-specific technologies
Optimizing AI-first workflows for professional delivery
Challengers work with large-scale systems, adapting and deploying fine-tuned models within enterprise constraints.
Fine-tuning frameworks, enterprise debugging and deployment tools
Integrating AI into production-scale enterprise systems
Modern AI meets legacy software as challengers coordinate agent teams to refactor, extend, and modernize complex codebases.
Agent orchestration frameworks, evaluation and debugging tools
Using agent teams to manage and reduce system complexity
Challengers rapidly learn new domains using AI, combining text, image, audio, and interactive systems into complete experiences.
Eleven Labs, Replicate, Midjourney, Stable Diffusion
AI-accelerated learning and cross-modal system integration
Focus shifts to designing scalable AI systems while beginning capstone development.
Cloud platforms, protocol implementations, observability systems
Production AI system architecture and platform design
Challengers complete and deploy a comprehensive AI system capable of learning, adaptation, and real-world optimization.
RL frameworks, full AI deployment pipelines
End-to-end frontier AI system delivery and presentation
Fellowship bonds
Is This You?
You need to be more than curious. You need to be committed.
You've shipped production systems — not just side projects
You're obsessed with getting better, not just getting by
You want feedback that's honest, not comfortable
You'll go all-in for ten weeks — no hedging
You'd rather be judged on what you build than where you went to school
Choose Your Path
Where you start depends on your experience. Where you end up depends entirely on you.
For engineers who want to work for companies
For engineers who want to work for the US Government
Upcoming Cohorts
Applications are reviewed on a rolling basis. Earlier is better. Waiting costs you.
Cohort 5
Starts April 27, 2026
Cohort 6
Starts July 6, 2026
Cohort 7
Starts September 14, 2026
The window is open
Ten weeks from now, you're either the same engineer — or a fundamentally different one.
You've read this far. You're not browsing. You're deciding. Cohorts fill fast. Don't wait.