Research Engineer - Decentralized Training and Inference Verification
Core
Design and ship statistical verification systems to ensure tokens and training contributions in a decentralized, trustless network match claimed models and parameters.
Role type
Senior IC research engineer (decentralized ML verification)
Builds
Inference verification service and training verification algorithms for a global GPU network
Domain
Decentralized machine learning, distributed systems, adversarial security
Deliverable
production ML models
Required skills
statistical verification, probability theory, experimental design, threat modeling, system design for untrusted compute, cryptographic proofs, fraud detection
Preferred skills
large-scale pre-training, RL post-training, decentralized ML security, adversarial threat models (poisoning, Sybil, collusion)
Technologies
statistical testing frameworks, GPU/Mac heterogeneous environments
Responsibilities
Enumerate and update threat models for malicious or careless workers; design and calibrate statistical tests with controlled error rates; build and run the verification service in the inference path
Seniority
Senior, hands-on IC
