Research Opportunities
Up to $100,000 for faculty-led research on how robots learn from people in the real world. Bring a sharp question about egocentric data, transfer or evaluation, and we'll help fund the answer.
Trace Labs builds data and infrastructure for real-world robot learning. We capture how people perform everyday physical work in real homes and workplaces through synchronized first-person video, motion, hand and body pose, depth and language annotation, and we build the perception and quality pipelines that make that data usable for training robots. Turning real-world activity into robot capability raises open questions no single team can answer alone. This program exists to support the academic research that addresses them.
We welcome theory, algorithms, empirical studies, data and evaluation methods, and research systems, both foundational and application-driven. A focused contribution to one well-posed question is enough.
Submit one PDF through the form below by the deadline, on US letter paper with 11-point font and 1-inch margins. No contact with Trace, letter of support or institutional nomination is required, and there's no limit on proposals per PI or institution.
Proposals are reviewed on a rolling basis as they arrive, and early submissions are strongly encouraged: funding decisions may be made before the deadline for particularly strong proposals. Proposals submitted after the deadline may still be considered, subject to remaining funds.
Collaborating teams designate one lead institution to receive the gift. Multi-institution arrangements are made by the institutions, not by Trace Labs.
Questions? Email research-partner@tracelabs.ai.
The research narrative is 3–4 pages, excluding references. We suggest this structure.
Project title; PI and co-PI names, titles and institutions; primary research theme (A–H); requested funding amount (USD); and project duration.
Concisely summarize the research problem, its significance, the proposed approach and key technical innovations, expected scientific contributions, and potential impact on robot learning and embodied intelligence.
Define the research problem and the fundamental scientific or technical challenges it addresses. Explain the limitations of existing approaches, why these challenges remain unresolved, and why solving them matters. Describe the project's relevance to the research themes, its potential to advance robot learning or embodied intelligence, its broader significance and long-term impact, and how the research could benefit Trace Labs.
State the primary objectives, hypotheses or research questions. Describe the central technical ideas, scientific insights and anticipated innovations. Discuss the most relevant prior work and its gaps, explain how the proposed research advances the state of the art, and identify what is fundamentally new and the key expected contributions.
Give a concrete, technically rigorous description of the proposed methods, algorithms, models, theoretical frameworks or experimental methodologies, with their underlying principles and rationale. Describe the anticipated technical difficulties and proposed solutions, with enough detail to assess originality, soundness and feasibility. Preliminary results may be included where available.
Present a rigorous, feasible plan to validate the hypotheses, approaches and expected contributions. Identify the benchmarks, datasets, platforms, baselines, metrics and comparisons you'll use, as applicable, and the experiments, ablations, analyses or human-subject evaluations appropriate to the work. Explain how results will show scientific advances and assess generalization and robustness, and discuss limitations, failure cases, and how unexpected or negative results would be interpreted.
Provide a research plan of up to 12 months with major activities, milestones and expected outcomes, plus key risks, dependencies and mitigations. Describe the team's relevant expertise and responsibilities, and the datasets, platforms, hardware and compute available. If Trace Labs data or other resources would help, specify the data types, modalities, approximate scale, annotation requirements, intended uses and any critical dependencies. Such resources are subject to availability and separate arrangements.
Thank you. Your submission ID is pending. We've sent a confirmation to ; keep it for your records. Proposals are reviewed on a rolling basis, and we'll be in touch by email.
