The role
Deepgram is hiring an LLM researcher to design experiments, drive distributed training runs and ship new models into production. The work centres on data curation, transformer optimization and RL training.
What you'll do
- Define new LLM research initiatives with other Research Staff
- Survey, evaluate and distill current methods from the literature
- Design and run experimental programs for LLMs
- Drive distributed transformer training jobs and deploy models to production
- Document and present results clearly to a target audience
What they're looking for
Must-haves
- 3+ years of applied deep learning research
- Proven LLM experience: data curation, distributed training, RL
- Strong Python and PyTorch coding
- Experience with auto-regressive and sequence-to-sequence transformers
- Experience with distributed computing and large-scale data processing
Nice to have
- Understanding of distributed training and inference schemes for LLMs
- Familiarity with RLHF labeling and training pipelines
- Published deep learning papers, especially on LLMs
The company
Deepgram
Voice AI platform with 200,000+ developers
Deepgram builds real-time speech-to-text, text-to-speech and voice agent APIs. More than 200,000 developers and 1,300 organizations build on its voice-native foundation models, and it has raised a Series C.
Summarised by Jobs That Travel from Deepgram's listing.
