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

📋 Description

• Develop and manage the end-to-end inference stack, covering pipeline-parallel execution, placement and routing, transport, serving engine, and failure management.

• Establish the technical vision for the inference stack and implement it effectively.

• Create, validate, and deploy innovative algorithms for rapid inference on consumer hardware across the public internet.

• Maintain a swift and dependable rollout pipeline for reinforcement-learning training.

• Construct the serving layer for models after they have been trained.

⛳️ Requirements

• Proven experience in shipping serving-engine internals or independently developing a large-scale inference system.

• Practical capability to undertake this work immediately.

• Published research in distributed inference or related fields, such as LLM serving systems, pipeline parallelism over slow networks, or decentralized training, or the ability to explain unpublished work.

• Experience with systems functioning in low-bandwidth, high-latency environments like the public internet.

• A conviction that Protocol Learning is a legitimate approach for collective, trustless, and sovereign AI.

• Familiarity with reinforcement learning post-training.

• Exposure to Apple silicon or MLX technologies.

• Knowledge of P2P networking and NAT traversal techniques.

• Experience in proprietary, open-weight, and open-source AI laboratories.

• Professional-level proficiency in English, both written and verbal.

• Ability to collaborate effectively across different time zones.