Member of Technical Staff, Research

Fireworks AI · Data AI · San Mateo, CA · Engineering

Research role focused on advancing LLMs and multimodal systems through foundational research, enhancing model efficiency, accuracy, and scalability for generative AI infrastructure. The role involves designing, implementing, and evaluating novel architectures and training methods, with a focus on transitioning research into production.

What you'd actually do

  1. Conduct foundational research to advance the capabilities, efficiency, and reliability of LLMs and multimodal systems
  2. Design, implement, and evaluate novel model architectures, training methods, and optimization techniques
  3. Collaborate with engineering teams to transition research prototypes into production-grade systems
  4. Analyze empirical results, identify performance bottlenecks, and iterate quickly to improve model quality
  5. Contribute to internal research strategy by identifying high-impact opportunities and emerging trends in AI

Skills

Required

  • Research background in Artificial Intelligence, Machine Learning, Physics, or similar field
  • Experience solving analytical problems using analytic and quantitative approaches
  • Experience communicating research to audiences with different backgrounds
  • Experience coding in C/C++, Python, or other similar languages

Nice to have

  • PhD degree in Computer Science, Computational Physics, Mathematics, or a similar field
  • Research and engineering experience demonstrated via grants, fellowships, patents, internships, work experience, and/or coding competitions
  • Experience having first-authored publications at peer-reviewed conferences or journals
  • Experience working with ML frameworks such as PyTorch, TensorFlow, or Jax

What the JD emphasized

  • first-authored publications at peer-reviewed conferences or journals

Other signals

  • foundational research
  • advance LLMs and multimodal systems
  • high-performance AI infrastructure
  • transition research prototypes into production-grade systems