> Markdown version of [/jobs/ext/4422183-founding-engineer-environments-and-data](https://www.wearedevelopers.com/jobs/ext/4422183-founding-engineer-environments-and-data). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Founding Engineer, Environments and Data

- **Company:** ONE STOP COLLECTIBLE CORP
- **Location:** San Francisco, CA, United States
- **Experience:** Starter
- **Salary:** $200,000.0 - $300,000.0
- **Contract:** Permanent contract
- **Skills:** Artificial Intelligence, Computer Programming, Data Pipelines
- **Published:** October 23, 2026
- **Apply:** https://startup.jobs/founding-engineer-environments-and-data-clera-10532233

## About the Role

* Hands-on research experience with synthetic data pipelines, benchmarks, evaluations, or post-training.
 * At least 1 to 2 years of experience at a fast-growing or AI-focused startup.
 * Experience building environments or benchmarks for frontier AI training or evaluation is valuable.
 * A track record of finding flaws in established benchmarks, improving them, and addressing reward hacking in evaluation systems.
 * Ability to simulate complex software systems that can withstand agent tool use.
 * Startup experience, strong performance in competitive mathematics, programming, debate, or games, and a collaborative approach under pressure are valued.

## Description

Build evaluation environments and benchmarks that measure whether AI agents can perform real company work and match or exceed human performance. This founding engineering role focuses on agent capability assessment and production readiness, helping establish reliable evidence of agent competence., * Design synthetic company environments with realistic emails, records, and files generated from scratch.

 * Build graders and evaluation systems that can assess agents even when their performance may exceed human reviewers' expertise.
 * Develop benchmarks whose rankings correlate with real-world production performance.
 * Find and address evaluation failure modes, including reward hacking and benchmark gaming.
 * Simulate tools such as inboxes and CRMs with enough fidelity to support agent interactions at scale.