Member of Technical Staff - AI/ML Infrastructure Engineer, RL Post-training

Preference Model • San Francisco, CA

5d

About The Position

Preference Model is a superintelligence data research company that builds reinforcement-learning environments for training capable, better-aligned superintelligences. The company focuses on improving AI training objectives and has built RL environments for several frontier labs. It is backed by $16 million in seed funding led by a16z. Its founding team has experience building data infrastructure and datasets at Anthropic.

The role builds infrastructure and systems that power frontier post-training research on large language models. The engineer will develop scalable compute, scheduling, and data infrastructure for research using in-house RL environments, while improving the systems and tools researchers use to run reproducible experiments. The work supports high-throughput research and helps reduce the time from idea to result.

The position collaborates closely with Research Engineers to turn research needs into infrastructure requirements and deliver improvements in response to their feedback. The posting refers to the role as a Senior ML Infrastructure Engineer, while the job title is listed as Member of Technical Staff - AI/ML Infrastructure Engineer, RL Post-training.

Requirements

• Strong software engineering fundamentals.
• Hands-on experience building production-grade LLM inference and training infrastructure, ideally from the ground up.
• Experience with LLM training and inference internals, such as transformers and distributed training, and inference libraries such as vLLM, SGLang, or Megatron.
• Experience with RL training frameworks such as Slime, veRL, Ray Train, SkyRL, or Miles.
• Significant experience and understanding of distributed systems principles.
• Hands-on experience with cloud platforms such as AWS or GCP and container orchestration such as Kubernetes, building systems for high-throughput, low-latency workloads.
• Experience with data engineering tools and building robust, scalable data pipelines.
• Proficiency with core ML frameworks such as PyTorch or JAX.
• Ability to balance production rigor with fast-moving research and clearly communicate infrastructure tradeoffs to researchers who are not infrastructure specialists.

Nice To Haves

• Experience building production-grade LLM inference and training infrastructure from the ground up is preferred.
• Experience with specific tools and frameworks including vLLM, SGLang, Megatron, Slime, veRL, Ray Train, SkyRL, or Miles.

Responsibilities

• Design, build, and scale compute, scheduling, and data infrastructure for post-training research on in-house RL environments.
• Develop and maintain core ML framework primitives and internal tools used by researchers, accelerating reproducible experimentation and reducing time from idea to result.
• Build evaluation and benchmarking infrastructure, monitoring, logging, debugging tools, and automated testing and deployment systems.
• Partner with Research Engineers to translate research needs into infrastructure requirements and ship improvements in response to their feedback.

Benefits

• Competitive cash and equity compensation (greater than the 90th percentile).
• Ownership and autonomy in a fast-moving startup environment.
• Opportunity to work alongside senior and staff engineers from frontier labs and infrastructure companies, as well as top ML engineers.
• Health, vision, and dental benefits.
• 401(k) match.
• Lunch provided onsite every day.
• Weekly snack orders.
• Visa sponsorship and relocation support available.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior
• Education Level: No Education Listed