Call for Proposals

Research Proposals for Faculty

JHU + Amazon AI2AI Initiative  |  2026–2027 Award Cycle

Abstract submission deadline extended: March 17

About This Call

AI2AI is a partnership between JHU and Amazon dedicated to bridging academic and industry research in artificial intelligence. With Amazon’s support, AI2AI annually funds multiple faculty research projects and doctoral fellowships. We invite proposals that take innovative approaches to a broad range of artificial intelligence topics.

Amazon builds and deploys AI across three interconnected technology layers: custom silicon and high-performance computing infrastructure; a broad Foundation Model selection encompassing both Amazon-built and third-party models; and generative AI applications and services that improve experiences for customers and employees alike. Three principles guide this work: a strategic focus on practical, real-world applications; world-class data, compute, and talent resources; and a firm commitment to responsible, reliable, and trustworthy AI.

2026–2027 proposals should advance research in one or more of the topic areas listed below. Proposers will select a primary and secondary category in the application. Feel free to bring your own unique viewpoint and expertise to these topics.

Topics of Interest

Foundation Model Improvements

  • Novel model architectures
  • Novel training algorithms and methodologies
  • Multi-modal (e.g., text, image, video, audio) understanding and generation
  • Multi-lingual understanding and generation
  • Algorithms and workflows for acquiring/curating high-quality and diverse datasets for training
  • Continual learning
  • RL methodologies including verifiable and unverifiable rewards

Foundation Model / Agentic Evaluation

  • Creation of new benchmarks for assessing foundation model capabilities
  • Methodologies for robust evaluation of generative AI systems, including agents

Efficient Generative AI & Efficient Inference

  • Efficient training and inference for cloud and on-device applications
  • Compute and memory efficient handling of multi-modal long/infinite context
  • Improving efficiency of models, including diffusion models
  • Novel and efficient approaches for inference-time scaling
  • Novel methods for accelerating LLM inference (speculative sampling and beyond)

Reasoning

  • Commonsense and domain-specific (e.g., math, coding) reasoning
  • Temporal and spatial reasoning
  • Reasoning for planning

Knowledge Grounding

  • Ground (multi-modal) generation on up-to-date world, domain-specific, enterprise, or personal knowledge
  • Memory-augmented generative AI systems
  • Efficient/accurate multimodal retrieval across heterogeneous sources (KGs, People Graphs, DBs, web, local search)
  • Blending and fusion of multiple knowledge sources for effective grounding
  • Improved interaction between LLMs and external knowledge sources

Agentic AI

  • Autonomous systems capable of performing tasks, making decisions, and interacting with environments/humans
  • Informal reasoning
  • Low-code/no-code for business agentic applications
  • Multi-agent systems and agent orchestration framework improvements
  • Customization and continual improvement of agents post deployment
  • Agent-environment simulation research and world models
  • Internationalizing agentic systems
  • Applications including Agentic Coding, Deep Research Agents, and Cybersecurity

Personalization

  • System-level personalization
  • Personalization of dialog-based applications
  • Prompt-based personalization of large language models
  • Personalized retrieval and ranking for RAG-LLM systems

Responsible Generative AI

  • Red teaming (including advanced approaches; for multi-modal models)
  • Improve RAI performance (robustness to jailbreaking / membership inference; watermarking; deepfake detection)
  • Responsible agentic AI (multi-agent robustness; adherence to guardrails)
  • Measurement/alignment against frontier risks (scheming, deception, AI autonomy)
  • International/cultural alignment

AI Accelerated Science and Engineering Innovation

  • Systems using Generative AI for advancing science & tech (physics, math, chemistry, biology, hardware, materials, engineering, economics, healthcare, climate)

Eligibility

Full-time tenure-track, research-track, and teaching faculty members at JHU are eligible to submit proposals as principal investigators. Collaborative proposals led by WSE faculty with colleagues from other JHU divisions are welcome.

Amazon Scholar Note: Faculty who will be Amazon Scholars in AY 2026–2027 are eligible to submit but must adhere to JHU Conflict of Interest policies.

Award Types

Sponsored Research Awards
Gift-Funded Awards

Terms comparable to other sponsored projects at JHU.

Expected to involve direct collaboration with Amazon researchers.

1–3 awards of up to $100,000 in direct costs anticipated.

Indirect costs covered; recipients will not be asked to reduce budgets to accommodate.

Treated as unrestricted gifts.

Intended to support exploratory, investigator-driven research.

2–4 awards of up to $100,000 each anticipated.

Applicable indirect charges will be added to awarded proposals.

The application process is the same for both award types. Award type will be determined in consultation with Amazon following selection.

Important Dates

Date Milestone
February 24, 2026 Portal opens — submissions accepted
March 17, 2026 Abstract proposals due (1-page max, via Amazon Research Portal)
April 6, 2026 Amazon reviews complete — invitations to submit full proposals issued
April 13–24, 2026 Optional research discussion and feedback sessions with Amazon scientists
April 27, 2026 May 5, 2026 Full proposals due (invited applicants only, via Amazon Research Portal)
June 30, 2026 Award decisions communicated to all applicants
August/September 2026 Award period begins (nominally September 1, 2026 – August 31, 2027)
Abstract submission is required in order to be invited to submit a full proposal. Only applicants whose abstracts are selected by Amazon will be asked to submit a full proposal. All abstract submitters will receive a response by April 6, 2026.

How to Apply

Step 1: Submit Your Abstract (Due March 17, 2026)

Abstracts are submitted through the Amazon Academic Research Portal. This is a new submission platform for the 2026–2027 cycle. Complete the steps below to register and submit:

1 Register Create an account at the Amazon Academic Research Portal.
2 Set Up Your Profile Select Johns Hopkins University as your institution and choose "Faculty" or "Research Scientist" as your Professional Role to access open Calls for Proposals.
3 Submit New Proposal Click "Submit New Proposal" in the top banner and select the AI2AI CFP.
4 Review Requirements Carefully read the submission requirements, instructions, and deadlines.
5 Prepare & Submit Complete the abstract details form and upload your 1-page abstract (maximum, excluding references).
Portal URL: https://home.academic-research-portal.amazon.science/login

Your abstract should be a concise overview of a proposed research project — one page maximum, excluding references. No budget or full proposal is required at this stage.

Step 2: Optional Feedback Session (April 13–24, 2026)

Amazon will host optional research discussion and feedback sessions for invited applicants between April 13 and 24. These sessions pair PIs with matched Amazon scientists to discuss research focus and proposal scope — we strongly encourage all invited applicants to participate.

Step 3: Submit Your Full Proposal (Due April 27, 2026 May 5, 2026)

Full proposals are by invitation only. Invitations and full proposal instructions will be issued by Amazon on April 6, 2026. Proposals should be submitted via the Amazon Research Portal by April 27, 2026 May 5, 2026.

Component Format / Length Notes
Project Overview PDF, 3–4 pages maximum Technical description, expected deliverables/outcomes, and milestones
Budget & Justification Excel, using the provided AI2AI budget template Include all direct costs; indirect charges added to awarded proposals
References No page limit —
PI Biographies Up to 3 pages per PI NSF biographical sketch format

Proposals currently under consideration by the Amazon Research Awards program should be noted to indicate concurrent submission. Proposals will not be selected for funding from both sources.

Funding Guidance

Projects are typically funded at up to $100,000 in direct costs, depending on project scope and nature of funding. If pursued as Sponsored Research, the AI2AI Advisory Board will discuss adjustments with the proposer and ensure applicable indirect costs are covered.

No proposal may contain confidential information or be marked as such. Please focus on objectives achievable without access to non-public Amazon data and scope proposals to use publicly available data or data that can be independently collected or generated.

Questions

For questions about this call, please contact ai2ai@jhu.edu or reach out to the Amazon program contact, Kathleen Allen (allenkdh@amazon.com).

AI2AI website: https://ai2ai.engineering.jhu.edu/opportunities/