# Foundry: Building the Foundational AI Stack for a Browser-Agent Future

## Foundry: Building the Foundational AI Stack for a Browser-Agent Future

Foundry (YC F24) is on a mission to move beyond superficial AI applications and build the fundamental infrastructure required for truly reliable and scalable browser agents. The company, founded by ex-Scale AI & Meta engineers and seed-funded, believes the future of automation lies not in API integrations, but in AI capable of navigating the complexities of the web. They’re seeking a founding engineer to help them realize this vision.

Currently, browser agents are akin to the early days of GPT – promising demos but ultimately brittle and unreliable in real-world scenarios. Foundry aims to bridge this gap by building a “world model” for the web, a foundational AI stack designed specifically for browser agents. This infrastructure includes:

* **Realistic web simulation:** A robust environment capable of modeling complex tasks, ambiguous user interfaces, and non-deterministic interactions.
* **Robust RL training loops:** Effective learning systems that can learn from sparse rewards and authentic user behavior.
* **Rigorous evaluation benchmarks:** Metrics tied to real-world outcomes and genuine business impact.

Instead of focusing on incremental tools or GPT wrappers, Foundry is building this core infrastructure from the ground up, envisioning it as the essential foundation for all future browser-agent development.

**The Opportunity: Founding Engineer**

Foundry is searching for a talented and driven founding engineer who is eager to build core ML systems and RL infrastructure from scratch. This role offers a unique opportunity to work at the cutting edge of AI and shape the future of web automation.

**Key Responsibilities:**

* Build a browser agent gym, tackling the challenges of sparse rewards, non-determinism, and messy DOM trees.
* Solve hard ML problems at the edge of production, on a daily basis.

**Ideal Candidate:**

* Killer software engineer (Python, TypeScript) with 0-2 years of experience.
* A proven track record of shipping code quickly, demonstrated through internships, side projects, OSS contributions, or research.
* Real-world ML exposure, including training models, building infrastructure, and running experiments.
* Strong OSS contributions or recognized publications.
* A strong desire to grind, improve, and build something truly novel.

**Bonus Points:**

Experience building RL agents, browser automation (Puppeteer, Playwright), or eval tooling is a plus. The ideal candidate is also someone who is tired of wrapping APIs and fine-tuning HuggingFace models and ready to delve into deeper, more impactful work.

**Why Join Foundry?**

Foundry offers the opportunity to work on real-world ML infrastructure with real users and tackle genuine technical complexity. The company provides early equity, competitive pay, and a massive growth trajectory. This is a fully in-person role, based in San Francisco, with a 5-day/week commitment.

If you’re passionate about building the foundational AI stack for a browser-agent future and are ready to move beyond superficial applications, Foundry wants to hear from you. Reach out directly and let’s build the future of web automation together.