Senior / Staff Machine Learning Research Scientist, Agents

Scale AI Scale AI · Data AI · San Francisco, CA · Research

Research Scientist role focused on building state-of-the-art AI agents, studying essential data types for agents like browser and SWE agents, and guiding data strategy to advance intelligent, adaptable AI agents. The role involves contributing to research publications, collaborating with customer researchers, and translating advancements into scalable solutions.

What you'd actually do

  1. study the data types essential for building state-of-the-art agents, such as browser and SWE agents
  2. explore the data landscape needed to advance intelligent, adaptable AI agents, guiding the data strategy at Scale to drive innovation
  3. contribute to impactful research publications on agents
  4. collaborate with customer researchers
  5. work alongside the engineering team to translate these advancements into real-world, scalable solutions

Skills

Required

  • LLMs
  • Pytorch
  • Jax
  • Tensorflow
  • interpreting research literature
  • ML prototyping
  • debugging
  • research concepts

Nice to have

  • open source LLM fine-tuning
  • bespoke LLM fine-tuning projects
  • agent frameworks (OpenHands, Swarm, LangGraph)
  • agentic reasoning methods (STaR, PLANSEARCH)
  • cloud technology stack (AWS or GCP)
  • developing machine learning models in a cloud environment

What the JD emphasized

  • track record of published research in top ML venues
  • practical experience working with LLMs
  • proficiency in frameworks like Pytorch, Jax, or Tensorflow
  • adept at interpreting research literature and quickly turning new ideas into prototypes
  • at least three years of experience addressing sophisticated ML problems

Other signals

  • research on agent environments
  • benchmarking autonomous agent performance
  • data programs to improve LLMs agentic capabilities
  • foundational tools and frameworks for evaluating models as agents
  • autonomous agents that dynamically interact with diverse external environments