Research Engineer - Decentralized Training and Inference Verification

What you'll do

• Maintain the threat model for malicious or careless workers across training and inference.
• Design, benchmark, and calibrate statistical verification tests with controlled error rates across heterogeneous hardware.
• Build and operate the verification service in the inference path.

What they ask for

• Experience building or publishing a calibrated statistical decision system with stated error rates.
• Deep expertise in statistics and probability, including experimental design and threshold calibration.
• Knowledge of verification approaches for untrusted compute, including statistical testing, re-execution, cryptographic proofs, and trusted hardware.
• Professional-level English proficiency, written and spoken.

Nice to have

• Familiarity with large-scale pre-training and RL post-training.
• Familiarity with decentralized ML security and adversarial threat models, such as poisoning, Sybil, collusion, and replay.

Tools

GPUs, Macs

AI focus

Verification for decentralized AI training and inference, including detecting incorrect model use, dishonest training contributions, poisoning, and data extraction.