Research Staff, LLMs

$150k - $250k • Remote • USA - Remote • FullTime

Posted 1mo ago

Job Location

USA - Remote

Tech Stack

Python PyTorch Reinforcement Learning Deep Learning Transformers RLHF Distributed Training LLMs optimization Transformer architecture data curation auto-regressive

Remote Work Policy

Fully remote

Employment Type

FullTime

About the job

Deepgram is a leading platform in the Voice AI economy, offering real-time APIs for speech-to-text, text-to-speech, and building voice agents. We are seeking an experienced researcher with extensive experience in Large Language Models (LLMs) and a deep understanding of transformer architecture to join our Research Staff. This role involves tackling the hard technical aspects of LLMs, including data curation, distributed large-scale training, transformer architecture optimization, and Reinforcement Learning (RL) training. We are looking for individuals who see unsolved problems as opportunities, can quickly validate ideas, scale proofs-of-concept, and are obsessed with using AI to amplify their impact. If you are energized by these challenges and driven to create elegant, scalable solutions, this role offers transformative potential.

Responsibilities

Brainstorm and collaborate on new LLM research initiatives.
Survey literature, evaluate, and distill current methods.
Design and execute experimental programs for LLMs.
Drive transformer (LLM) training jobs on distributed compute infrastructure and deploy new models.
Document and present research results and technical concepts clearly.
Stay updated on deep learning and LLM advancements and their product implications.

Requirements

3+ years of experience in applied deep learning research.
Solid understanding of neural network types, architectures, and loss mechanisms.
Proven experience with large language models (LLMs), including data curation, distributed large-scale training, transformer architecture optimization, and RL learning.
Strong coding experience in Python.
Experience working with PyTorch.
Experience with various transformer architectures (e.g., auto-regressive, sequence-to-sequence).
Experience with distributed computing and large-scale data processing.
Prior experience conducting experimental programs and optimizing models based on results.
Deep understanding of transformers, causal LMs, and their underlying architecture.
Understanding of distributed training and inference schemes for LLMs.

About Deepgram

View company profile