Sepal AI is working with top-tier AI research labs (OpenAI, Anthropic, Google) to ensure next-generation AI systems behave safely and reliably inside real-world enterprise software. We’re building safety benchmarks that simulate business workflows—like Salesforce—and uncovering how AI models might help or fail in high-
About the role
Sepal AI is working with top-tier AI research labs (OpenAI, Anthropic, Google) to ensure next-generation AI systems behave safely and reliably inside real-world enterprise software. We’re building safety benchmarks that simulate business workflows—like Salesforce—and uncovering how AI models might help or fail in high-stakes environments.
🧠 What You'll Do
- Recreate Salesforce workflows (e.g., Lead-to-Cash, CPQ) to test AI-assistant behavior.
- Design realistic tasks and edge cases where AI could make risky or costly errors.
- Create structured rubrics to evaluate whether AI decisions are safe, compliant, and reliable.
- Collaborate with AI researchers to refine safety tests and improve model alignment with enterprise needs.
✅ Who You Are
- 3+ years configuring or analyzing Salesforce systems.
- Strong with declarative Salesforce tools (Flows, Validation Rules, Reports).
- Familiar with risk-prone workflows like quoting, deal desk, or approvals.
- Bonus: background in QA, process documentation, or internal audits.
💸 Pay: $30–60/hr
🌍 Remote
⏱ Flexible hours
🧭 Impact: Your expertise defines the standards AI must meet before it can be trusted in critical business workflows.
Official source: expert-hub.sepalai.com. Details can change; the platform's page is the final word. How we check listings.
