jobs
Research Intern, Frontier Speculative Decoding (Winter 2027)
About the Role The Speculative Decoding team at Together AI builds the draft models and algorithms that let large language models generate faster without changing their outputs.
En anglais
- Company
- Together AI
- Location
- San Francisco
- Status
- Closed
- Posted
- 2026-10-16 22:48 UTC
An internship on Together AI's Speculative Decoding team, which builds draft models and algorithms that speed up large language model generation without changing outputs. The intern will prototype decoding methods and drafter architectures, train or distill draft models on dedicated GPU clusters, test them on realistic serving workloads, and help bring successful ideas into serving engines. The posting seeks a master's or PhD student in computer science, electrical engineering or a related field, with deep learning fundamentals and distributed training experience. It runs 12 to 14 weeks in San Francisco.
Original job posting