Senior Research Scientist - Reinforcement Learning, MoEs(m/w/x)

Description

You will push the frontiers of AI by developing advanced agentic systems and MoE models, transforming complex reinforcement learning research into reliable features for millions of users.

Requirements

• •Expertise in MoEs, LLMs, VLMs, and Diffusion
• •Experience modifying and adapting open-source models
• •Strong experimental design and reproducibility skills
• •Fluency in Python and PyTorch
• •Experience building and evaluating agent loops
• •Hands-on experience with policy optimization
• •Experience with large-scale distributed training
• •Experience with RL for MoE architectures
• •Experience with video and audio modelling
• •Experience with multi-agent settings
• •Strength in alignment and safety evaluations
• •Contributions to open-source or benchmarks

Tasks

• •Develop multimodal agentic systems
• •Design reward models and learning loops
• •Scale reinforcement learning across distributed systems
• •Optimize mixture-of-experts architectures
• •Build scalable training and evaluation loops
• •Construct datasets for post-training
• •Design experiments for policy optimization
• •Develop simulation and sandbox tasks
• •Identify failure modes like hallucinations
• •Implement offline suites and A/B tests
• •Collaborate with product and platform teams
• •Mentor teammates and share research findings

Tools & Technologies

Benefits