About the job
All candidates are expected to work from our Seattle offices, with specific on-site requirements varying by position and team. For inquiries regarding on-site arrangements, please consult your recruiter.
Our base salary range is $128,880 - $193,320, complemented by a competitive bonus structure to enhance your overall compensation.
Who You Are:
The Allen Institute is on the lookout for a skilled and driven Research Engineer II to join our FlexOlmo team, where you will focus on developing large language models optimized for flexible data utilization, with a strong emphasis on Mixture-of-Experts (MoE), long-context language models (LCLMs), and retrieval systems.
You are an innovative engineer who excels in a dynamic atmosphere, is self-motivated, and thrives as a collaborative team member. Your profound understanding of contemporary deep learning techniques, natural language processing, and transformer architecture, particularly MoEs, allows you to transform high-level objectives into actionable research and implementation strategies. You possess the ability to articulate complex technical concepts clearly, contributing to real-world advancements for researchers and practitioners alike. Your enthusiasm for creating leading-edge open AI models is palpable.
Who We Are:
As a non-profit AI institute, our mission is to spearhead foundational AI research and innovation that yields tangible impacts through expansive open models, data, and artifacts such as OLMo, Tulu, Asta, and OlmoEarth. We bring together exceptional scientific and engineering talent to explore the vast potential of open AI. Our collaborative efforts allow us to swiftly identify and act upon the most exciting and promising ideas in AI.
The FlexOlmo team is dedicated to designing new architectures and training methodologies that enhance models' data utilization, improving training, inference-time conditioning, and retrieval capabilities to broaden data types and boost performance. We prioritize developing scientific methodologies for evaluating and comprehending these systems, producing high-impact research and expertly engineered open-source tools that foster NLP research globally. Our first release in July 2025 will focus on new Mixture-of-Experts models.
