Research Internship, AI for Science: The Allen Institute for Artificial Intelligence (Ai2)

This 12-week internship with the Asta Research Team (which includes Semantic Scholar and Aristo) focuses on rapidly advancing science using intelligent and interactive AI research assistants. The team is interdisciplinary, working on AI, HCI, ML, and NLP. Interns will conduct high-impact research, author scientific papers for top-tier conferences, and collaborate with Ai2 scientists and engineers.

Focus Areas
Agentic planning, reasoning, learning, and evaluation.

Automated and AI-assisted scientific discovery.

Data-driven discovery.

Human-AI Collaboration (Literature Sensemaking, Research Ideation).

Human-Agent Collaboration: Proactivity, Long-running agent, Memory and context management.

Scientific Literature Understanding, RAG, Multi-doc summarization.

Continual learning, representation, and reasoning.

Post-training, model tuning, distillation, and specialization.

Responsibilities
Conduct high-impact research and author scientific papers for publication in high-profile computing conferences (e.g., ACL, NeurIPS, CHI) and journals.

Collaborate with and learn from a dedicated Ai2 mentor and other team members.

Requirements
PhD Student: Completing a PhD with research experience in Computer Science or a related field.

Research Focus: Specific focus on one or more of the team’s active research areas.

Publication Record: A record in AI-related venues (e.g., ACL, EMNLP, NeurIPS, ICML).

Software Skills: Strong software engineering skills. Experience with system building (e.g., React, Typescript), deep learning frameworks (e.g., PyTorch), or agentic frameworks (e.g., Autogen).

Benefits
Medical, dental, and vision coverage.

Commuter/internet stipend ($125/month) and fitness/wellbeing stipend ($200/month).

Accrued sick days and paid holidays.