Research Engineer, Asta

The Allen Institute for Artificial Intelligence•Seattle, WA

1d•Onsite

About The Position

Persons in these roles are expected to work from our offices in Seattle. On-site requirements vary based on position and team. If you have questions about on-site work arrangements for this role, please ask your recruiter. Our base salary range is $118,800 - $178,200, and in addition we have generous bonus plans to provide a competitive compensation package. Are you passionate about AI and its potential to help accelerate science? Ai2's Asta Project aims to rapidly advance science, developing solutions in the areas of AI research assistants for scientists, literature understanding and synthesis, end-to-end discovery, human-AI interaction and intelligent interfaces, data-driven discovery, agents that can learn over time, methods for training agentic models, and the evaluation and benchmarking of agents. We are looking for a talented and motivated Research Engineer, ready to take on a complex challenge with huge potential impact for the future of AI, in particular, with skills and interest in the following focus areas: Automated and AI-assisted scientific discovery. Agentic planning, reasoning, learning, and evaluation. Data-driven discovery. Literature-driven ideation and discovery. Human-Agent Collaboration: Proactivity, Long-running agent, Memory, and context management. Continual learning, representation, and reasoning. Post-training, model tuning, distillation, and specialization. The Asta Team (Semantic Scholar and Aristo, based in Seattle) at the Allen Institute for AI (Ai2) is an interdisciplinary research and engineering team focused on AI, ML, NLP, and HCI to support our mission of accelerating science with intelligent and interactive research assistants. All of our scientific artifacts (data, model, software, papers) are open-sourced and maintained. We regularly publish and present at high-profile journal/conferences in ML, NLP, HCI, and Accessibility.

Requirements

A bachelor’s degree in CS/EE/Data Science/Applied Mathematics/Statistics/ML/NLP, or a related field, or equivalent relevant experience, and expertise in building ML infrastructure.
2+ years of experience building agentic infrastructure that handles tools, skills, and other artifacts.
2+ years of experience building infrastructure that handles data preprocessing/transformation and machine learning model training, evaluation, inference, and deployment.
Knowledge of modern deep learning, natural language processing, and reinforcement learning techniques.
Strong software engineering skills, particularly around building performant systems and debugging.
Must have experience with Python and PyTorch/Jax/Tensorflow, agentic frameworks (e.g., MCP), as well as feel at ease in picking up new programming languages, libraries, or APIs as tools as project needs evolve.
Familiarity with cloud compute resources (e.g., GCP, AWS, Modal) and containerization (e.g., Docker).
Strong collaboration and communication skills - our environment is small and collaborative, and we'd like you to thrive while working closely with others, sometimes with complementary skills/perspectives.

Nice To Haves

Advanced degree in CS/EE/Data Science/Applied Mathematics/Statistics/ML/NLP or related fields and/or relevant and equivalent engineering experience.
Contributions to open-source ML or research libraries (e.g., spaCy, AllenNLP, transformers, langchain).
Experience successfully operating at scale in a production setting.
Experience in HPC settings.
Curiosity about AI research.

Responsibilities

Building infrastructure to facilitate the next generation of LLM and agentic research.
Creating AI tools to facilitate scientific discovery in domains such as biology, cancer research, neuroscience, social science, etc.
Designing, building, and training machine learning or language models for agentic workflows.
Bridging the gap between cutting-edge research and a widely adopted product.
Bringing software engineering best practices to a research environment.
Supporting and collaborating with an open-source community.
Releasing your contributions back to the broader community in the form of open source software, model releases, and additions to Ai2’s public API and open research datasets, as well as technical reports.