Every project on this site started with someone saying “we still do that by hand.” The first step is a conversation with Dave and Alex. No scope, no commitment.
Thirty minutes on what is slow, manual or undocumented in your operation. We will tell you honestly whether AI is the right tool.
We write a statement of work together and you name a champion, usually a president, CEO or VP, who owns the engagement.
Two to three fellows work the project weekly, with a project coach clearing blockers and CRC advisors supporting the heaviest builds.
A live demo, a documentation package and a working system. You decide whether it continues into the academic year.
Not every good business problem makes a good student project. These are the questions we work through together, and they are meant to open up possibilities rather than rule things out.
What is the challenge, and what would success look like? The more useful question is usually why it has not been solved yet.
Fellows cannot build against a system they cannot see. What exists matters more than whether it is tidy.
The most valuable person understands the problem deeply. They do not need to know what should be built. That is the project.
Worth deciding early, because it shapes the scope.
None of this is a gate. It is a way to think out loud about what a project could be, and a conversation is usually faster than a form.
Four roles on every engagement. The combination is what makes this different from an internship.
The client. A president, CEO, or VP who owns the engagement and can unblock data access in days rather than weeks.
The engagement manager from iNDustry Labs. Sets cadence, runs check-ins, owns scope conversations with the champion.
Two to three students who do the work, drawn from different majors and class years by design.
Technical bench from Notre Dame's Center for Research Computing, backing the heaviest builds.
Each project ran twelve weeks with a named champion at the company, an iNDustry Labs project coach, and technical advisors from the Center for Research Computing. Published with company approval.
Filtered by business function rather than industry, the same bottleneck tends to look similar across sectors.
A text-to-SQL agent over PLEX ERP and Paperless Parts data, with a benchmark framework built to catch hallucinated joins before a plant manager ever sees a wrong number.
A pipeline that reads decades of engineering drawings and rebuilds them as parametric SOLIDWORKS feature trees, preserving the original design intent rather than just the geometry.
Government quoting demands ITAR, NSN, and ASTM compliance across hundreds of line items. One agent parses the RFQ, one sources qualified suppliers, one scores what comes back.
Twelve thousand employees across a thousand job codes, each carrying SOC, EEO-1, FLSA, and Workers' Comp metadata that cannot be lost in a merge, plus an agent that maps new roles to the consolidated set.
Auto-generating 2D drawings from 3D models inside Solid Edge, capturing the datum-handling rules that live in senior engineers' heads rather than in any documentation.
A retrieval agent across SAE, IEEE, internal, and customer specs, built with strict citation rigor, because an engineer who cannot verify a clause reference will not use the tool twice.
A prioritization agent wired into Microsoft Planner that surfaces critical-path tasks and path blockers, so leadership attention follows the schedule rather than whoever asked most recently.
An agent that validates webform leads against web research and prior interaction history, then routes qualified ones into Sugar CRM along with the NDA exchange workflow.
Candidates decline offers because they don't know enough about the region to say yes. A cost-of-living comparator and benefits view built to be shared during active recruiting.
PLACEHOLDER. A showcase recording exists for this project; project detail still to be confirmed with Dave.
PLACEHOLDER. A showcase recording exists in the Drive folder, but no other reference to this engagement was found. Needs confirming with Dave.
No projects in that function yet.
PLACEHOLDER: cost model to confirm.
A standing weekly call with the champion plus a day-to-day contact. The heaviest lift is front-loaded: getting fellows access to the systems and data they need.
You get a live demo, documentation, and the working system. Some projects continue into the academic year. Three of this summer's thirteen were continuations.
No. This is an educational fellowship inside a university. Projects are chosen partly for what they teach. The output is real, but the purpose is students.
The fellow team, the project coach, and CRC advisors where needed. PLACEHOLDER: formal data handling and NDA language to confirm.
Only with your approval. Nothing appears on this site without sign-off from your champion, and we never publish financial detail.
PLACEHOLDER: IP terms to confirm and state plainly here.
Repetitive knowledge work with a clear input and output: quoting, document conversion, lead qualification, or querying a system that currently requires a specialist. If someone does it the same way fifty times a week, it is worth a conversation.
If there is no project here, we will tell you.
Trouble with the embed? Open the scheduler in a new tab.
