Senior Research Engineer, Olmo + Molmo

Sourced from feeny.ai

Ai2 is hiring a Senior Research Engineer to train and advance its flagship open models, including Olmo and Molmo. You will own end-to-end model development, from building large-scale training and inference infrastructure to designing and evaluating multimodal and agentic systems. The role blends rigorous open research with strong software engineering practices and open-source releases.

• Build and optimize infrastructure for LLM, multimodal, and agentic research, including training/inference pipelines, dataset curation, and large-scale preprocessing
• Design, train, and evaluate multimodal models and agentic workflows covering tool use, planning, and long-horizon tasks
• Scope and lead research projects, prioritizing experiments for highest impact
• Apply strong software engineering practices to a research environment and bridge research to production-quality products
• Support the open-source community through model releases, datasets, public APIs, and technical reports

• 8 years of industry experience with 4+ years in ML infrastructure: data preprocessing, model training, evaluation, inference, and deployment
• Experience with end-to-end model development: dataset construction, training, fine-tuning, evaluation, profiling, and monitoring
• Familiarity with modern architectures including LLMs (MoEs, long-context) and vision-language models (e.g., Molmo, LLaVA)
• Knowledge of agentic systems: tools, memory, and long-running workflows
• Strong software engineering fundamentals and confident debugging of performant, scalable systems
• Proficiency in Python and a major ML framework (PyTorch, JAX, or TensorFlow)
• Familiarity with cloud and containerization (e.g., GCP, AWS, Docker)
• Strong communication and collaboration skills
• BS or MSc in Computer Science, Statistics, Engineering, Applied Mathematics, or a related quantitative field (or equivalent experience)

• PhD in machine learning or equivalent hands-on industry research experience
• Prior experience training and releasing large open foundation models