Prompt Engineer

About us
UnboundCompute is building an autonomous security researcher for web apps and APIs. Instead of matching known payloads, it learns how an app is meant to work, forms ideas about where the logic breaks, runs experiments, and proves real findings with hard evidence. We are early and building in the open.

The role
You will shape how frontier models reason about unfamiliar applications and turn that reasoning into verified security findings. This is core product work, not prompt tweaking on the side. You will design the prompts and workflows that drive the whole loop: understand, assume, experiment, verify, chain.

What you will do

• Design and refine prompts and agent workflows that find and prove real vulnerabilities
• Build evals that measure whether the system actually catches bugs, not just sounds smart
• Turn messy model behavior into reliable, repeatable steps
• Work closely with founders on the methodology itself

What we look for

• You have gotten real, non obvious results out of large language models
• You think in experiments and care about evidence
• Comfort with web and API basics; curiosity about how apps break is a big plus
• Clear writer. Most of the job is precise instructions.

Nice to have

• Security background (CTFs, bug bounty, appsec) or a track record of teaching yourself fast