[Hiring] Research Staff, LLMs @Deepgram

Role Description

Deepgram is currently looking for an experienced researcher who has worked extensively with Large Language Models (LLMs) and has a deep understanding of transformer architecture to join our Research Staff. As a Member of the Research Staff, this individual should have extensive experience working on the hard technical aspects of LLMs, such as:

• Data curation
• Distributed large-scale training
• Optimization of transformer architecture
• Reinforcement Learning (RL) training

Qualifications

• 3+ years of experience in applied deep learning research, with a solid understanding toward the applications and implications of different neural network types, architectures, and loss mechanisms
• Proven experience working with large language models (LLMs) - including experience with data curation, distributed large-scale training, optimization of transformer architecture, and RL Learning
• Strong experience coding in Python and working with Pytorch
• Experience with various transformer architectures (auto-regressive, sequence-to-sequence, etc.)
• Experience with distributed computing and large-scale data processing
• Prior experience in conducting experimental programs and using results to optimize models

Requirements

• Deep understanding of transformers, causal LMs, and their underlying architecture
• Understanding of distributed training and distributed inference schemes for LLMs
• Familiarity with RLHF labeling and training pipelines
• Up-to-date knowledge of recent LLM techniques and developments
• Published papers in Deep Learning Research, particularly related to LLMs and deep neural networks

Company Description

Deepgram is the leading platform underpinning the emerging trillion-dollar Voice AI economy, providing real-time APIs for speech-to-text (STT), text-to-speech (TTS), and building production-grade voice agents at scale. More than 200,000 developers and 1,300+ organizations build voice offerings that are ‘Powered by Deepgram’, including Twilio, Cloudflare, Sierra, Decagon, Vapi, Daily, Cresta, Granola, and Jack in the Box.

Deepgram’s voice-native foundation models are accessed through cloud APIs or as self-hosted and on-premises software, with unmatched accuracy, low latency, and cost efficiency. Backed by a recent Series C led by leading global investors and strategic partners, Deepgram has processed over 50,000 years of audio and transcribed more than 1 trillion words. There is no organization in the world that understands voice better than Deepgram.