Research Engineer – Post-Training

This is a fully remote position, open to applicants in United States, +1 more country.

📋 Description

• Develop the end-to-end RL post-training infrastructure, encompassing rollout ingestion, reward calculation, policy modifications, and the distribution of updated weights throughout the network.

• Establish the vision for the post-training stack and oversee its development.

• Modify RL algorithms to function effectively in asynchronous, high-latency, and partially trusted generation environments.

• Tackle issues related to staleness tolerance, off-policy adjustments, and communication-efficient policy enhancements.

• Create evaluations that showcase improvements in model performance.

• Release the first decentralized post-trained model as a publicly accessible artifact.

⛳️ Requirements

• Practical experience in executing RL post-training on large language models, such as RLHF, RLVR, or reasoning-oriented RL.

• Familiarity with rollout generation, asynchronous training processes, and weight synchronization.

• Proficient skills in Python and PyTorch at a production level.

• Experience with concurrency, failure management, and profiling prior to optimization.

• Publications in areas related to RL post-training, asynchronous or distributed RL, or similar fields, or unpublished work that can be thoroughly defended.

• A strong belief in Protocol Learning as a feasible approach for collective, trustless, and sovereign AI.

• Proficient in English, both written and spoken, at a professional level.

• Ability to collaborate effectively across different time zones.

• Nice to have: experience in training over slow networks or with decentralized/federated configurations.

• Nice to have: knowledge of vLLM or SGLang serving internals.

• Nice to have: experience with reward modeling or datasets that verify rewards.

• Nice to have: familiarity with P2P networking and NAT traversal.

• Nice to have: experience in proprietary, open-weight, and open-source AI laboratories.

🏝️ Benefits

• Significant equity ownership for key technical contributors, in addition to a competitive base salary.

• Flexible working environment with team members located globally.

• Optional full visa sponsorship and relocation assistance to either Australia or the US.