Research Staff, LLMs [EMEA - Remote]

About the role

Deepgram is a platform for real-time speech processing and voice AI products. This Research Staff position focuses on large language models and transformer-based systems, including data curation, distributed training, and RL training, to advance our LLM research initiatives.

What you'll do

• Brainstorm and define new LLM research initiatives with the Research Staff
• Survey literature, classify methods, and distill current approaches
• Design and execute experimental programs for LLMs
• Run transformer training jobs on distributed compute and deploy models to production
• Document and present results clearly for a technical audience
• Stay up to date with advances in deep learning and LLMs and assess their product implications

What they're looking for

• 3+ years of applied deep learning research with neural network expertise
• Proven experience with large language models including data curation, distributed training, transformer optimization, and RL
• Strong Python coding and PyTorch experience
• Experience with various transformer architectures and distributed computing
• Prior experience conducting experimental programs and using results to optimize models
• Nice to have deep understanding of transformers, causal LMs, distributed training/inference, and RLHF pipelines

For applicants in Nigeria

• The listing specifies EMEA remote; it does not state Nigeria-specific eligibility beyond remote work. Confirm timezone and eligibility before interviewing.
• Timezone expectations are not stated; confirm during interview process.
• The listing does not state contractor vs employee status; confirm during interview.
• Pay currency or range is not specified; confirm compensation details during interview.
• Visa or work-authorization notes are not specified; confirm eligibility requirements during interview.