Member of Technical Staff, Head of Quality at Plato

As the first Head of Quality, you will architect the automated verification systems and red-teaming suites that prevent AI agents from “gaming” training signals. You’ll own the final sign-off for environments shipped to frontier labs, ensuring RL data is grounded, robust, and free from reward hacking or brittle logic.

• Architect automated red-teaming suites and “LLM-as-a-judge” harnesses to stress-test task feasibility and aggressively eliminate reward hacking or task traps.
• Gate final delivery of high-fidelity simulations, auditing trajectories and reward dynamics to ensure models learn grounded behaviors rather than exploiting environment loopholes.
• Build and lead a high-agency verification team, scaling technical evaluation functions that catch synthetic drift and out-of-distribution behaviors upstream of model training.

• 3+ years of experience in software or ML engineering with an adversarial instinct for how optimizers exploit edges in reward functions or sandboxed environments.
• Proven ability to perform deep trajectory and code forensics, identifying subtle hallucinations or ungrounded assumptions within raw rollouts and reasoning traces.
• High-agency technical leader capable of building evaluation functions from zero to one, with the statistical depth to quantify quality and build custom AI-native tooling.

===

• Own the ground truth for frontier AI by serving as the final checkpoint before environments and datasets ship to world-leading research labs and hyperscalers.
• Join an elite, early-stage team backed by top-tier investors and researchers, working directly with founders Robert Farlow and Pranav Putta at the bleeding edge of RL.
• High-leverage impact as the first quality hire, building Plato.so’s automated judge harnesses and evaluation infrastructure from scratch with a $180k-$280k base salary.