Senior Applied Scientist

This is a fully remote position, open to applicants in United States.

📋 Description

• Take charge of LLM post-training processes, encompassing SFT, DPO, and reinforcement fine-tuning (RFT/GRPO), executing training from start to finish on GPU infrastructure.

• Develop and train rapid, multi-category reward models that generate scalar signals from contrastive preference pairs.

• Create on-policy and real-time assessments by scoring live agent trajectories utilizing LLM-as-a-Judge frameworks.

• Convert offline evaluation criteria into adaptable reward functions applicable to various production traces.

• Navigate complex and ambiguous tasks from inception to completion in partnership with product, engineering, science, platform, and data teams.

• Offer technical guidance and mentorship.

• Transform cutting-edge AI research into practical production outcomes.

⛳️ Requirements

• Master's degree or higher in Computer Science or a related discipline.

• Practical experience with LLM post-training and RL fine-tuning: SFT, DPO, RFT/GRPO, conducting end-to-end training on GPU infrastructure.

• Familiarity with reward model development; research-level experience is acceptable.

• Profound understanding of generative AI, including foundational models, transformers, reinforcement learning, and preference learning.

• Capability to independently define and resolve ambiguous challenges from start to finish while providing technical leadership to scientists and MLEs.

• Proficient programming abilities, particularly in Python.

• Experience with ML frameworks such as PyTorch or TensorFlow.

• PhD in Computer Science, Machine Learning, or a related field (preferred).

• Background in evaluating agentic AI: multi-turn trace analysis, LLM-as-a-Judge, and understanding the interplay between offline evaluations and online monitoring (preferred).

• Published research in post-training, RLHF/RLAIF, preference learning, or reward modeling (preferred).

• Familiarity with GPU training platforms such as Databricks or Fireworks (preferred).

🏝️ Benefits

• Equity awards determined by factors such as experience, performance, and location.

• Remote work from a location of your choice.

• Flexibility to work from wherever you are most productive.

• Support for accommodations related to disabilities or special needs.