Research Internship, Advancing Open Agentic LLMs at The Allen Institute for AI

Seattle, Washington, United States

The Allen Institute for AI Logo
$94,320 – $140,000Compensation
InternshipExperience Level
InternshipJob Type
YesVisa
AI ResearchIndustries

Requirements

  • Pursuing PhD (or, in exceptional cases, MS or BS) in computer science or a related field
  • Strong research background (for master/undergraduate students)
  • Demonstrated interest in advancing the state of the art in large language model (LLM) based agents and associated environments
  • Passionate, early-career researcher suitable for full-time 12-week internship
  • Ability to work from offices in Seattle (on-site requirements vary)

Responsibilities

  • Participate directly in Ai2's groundbreaking work, including building state-of-the-art AI systems
  • Develop challenging benchmarks
  • Conduct extensive experiments
  • Investigate fundamental questions in agentic LLMs, such as data synthesis, training objectives, learning environments, benchmarks, reasoning, tool use, external memory, safety, reliability, controllability, and theoretical understanding of agentic abilities
  • Work with a mentor as part of a team with world-class AI researchers and software engineers
  • Contribute to larger projects such as OLMo and Asta
  • Utilize Ai2’s compute resources (clusters of A100s and H100s)

Skills

LLMs
agentic models
data synthesis
training objectives
learning environments
planning
reasoning
tool use
APIs
code writing

The Allen Institute for AI

Advances AI through research and engineering

About The Allen Institute for AI

The Allen Institute for Artificial Intelligence (AI2) focuses on advancing artificial intelligence through research and engineering. It develops technologies that improve machine reasoning, common sense for AI, and language modeling. AI2 works on various projects, such as AllenNLP for natural language processing and EarthRanger for environmental monitoring. Unlike for-profit companies, AI2 operates as a non-profit, relying on grants and donations to fund its initiatives. This allows them to provide their research and tools to the public for free, promoting collaboration in the AI field. The goal of AI2 is to tackle significant scientific and engineering challenges to benefit society.

Seattle, WashingtonHeadquarters
2014Year Founded
N/ACompany Stage
AI & Machine LearningIndustries
201-500Employees

Benefits

Medical
401(k)
Visa Sponsorship
Time off
Catered meals & free snacks
Training & development
Tuition reimbursement
Remote & hybrid work options

Risks

Competition from OpenAI's QwQ-32B model may reduce AI2's market influence.
AI2's reliance on grants poses financial risks during economic downturns.
Open-source models may face intellectual property risks and misuse.

Differentiation

AI2 focuses on open-source models, promoting transparency and collaboration in AI development.
AI2's projects like AllenNLP and Semantic Scholar address diverse scientific and societal challenges.
AI2 operates as a non-profit, relying on grants and donations for funding.

Upsides

AI2's open-source models attract researchers, enhancing collaboration and innovation.
Tülu3 and OLMo-2 models position AI2 as leaders in open-source AI.
AI2's multimodal models meet the demand for integrated data processing in various fields.

Land your dream remote job 3x faster with AI