These 2 Builders Discovered the Potency of Going Analog: Reflections on the Human Power of CMU’s Startup Week

Carnegie Mellon University Startup Week brings thousands of people from the startup, research and investment worlds together to learn and swap ideas. Participants from last year’s inaugural event attest that the gathering across campus provided more than valuable information. It delivers invaluable connections through conversations.

Two founders — both with CMU ties — explained how that experience helped grow their businesses.

In the highly technical, complex world of machine learning, professor Pradeep Ravikumar(opens in new window), leader of the Neuro-Symbolic AI Group at CMU's Machine Learning Department(opens in new window) in the School of Computer Science(opens in new window), said the most important takeaway from CMU Startup Week(opens in new window) is the vibrancy of face-to-face conversations.

“The energy from other builders and the knowledge of investors is psychologically so important. It was just a great vibe,” said Ravikumar, who founded Wood Wide AI(opens in new window). The company solves a problem long faced by machine learning — finding important information in data before leaders know exactly what questions to ask. Unlike a typical dashboard, which is built to track specific, predetermined metrics, the product can uncover unexpected patterns and problems as they emerge.

For example, a retail company could use Wood Wide AI to identify an unexpected problem affecting sales. Imagine a national retailer notices that sales are dropping at several stores. A traditional dashboard might show that sales are down, but managers would still have to figure out why.

Wood Wide AI could analyze data from across the company and uncover an unexpected connection — for example, that stores experiencing the decline are all receiving late deliveries of a particular product because of a disruption at one distribution center. The company could spot the pattern early and take action before the problem spreads or becomes more costly.

At the time of the 2025 Startup Week, the company was in its early phases, an idea with much potential. Ravikumar was focusing on customer discovery and testing product iterations. Talking to fellow attendees and other founders, he realized he was not alone in some of his business-building challenges.

Since then, the venture transitioned from initial discovery to developing structured data intelligence platforms with design partners.

A dedicated teacher of machine learning at CMU, Ravikumar noted that in-person networking for builders provides crucial energy that just doesn’t happen in traditional academic environments.

“We learned about problems we wouldn’t have discovered on our own, which helped us focus our development on what people actually needed,” he said.

As a first-time founder, Ravikumar believes there is no substitute for talking directly with mentors and others who have been through the process. Those conversations can reveal insights and challenges that are difficult to find through internet research alone. He says a key takeaway was the mental boost realizing that so many others are out there trying, too.

“It spurs you to kind of try different directions, and that is a very underrated aspect of startups — the more psychological aspect, the more formational end of things,” Ravikumar said.

Last year, Ravikumar said they were still in early stages, cobbling together output with contractors, trying to get a sense of the market. By talking to investors at CMU Startup Week, he learned what he did not know, which he cites as a theme founders come up against.

“There was infrastructure in the space already, because the pain points companies have with this are to make good decisions based on data,” he said.

Since CMU Startup Week, he grew the company from one to eight staff members in a dedicated workspace. Ravikumar has shifted his focus from teaching to running the company full-time, stepping away from CMU responsibilities for a year.

“We can do things that we just dreamed about more like sci-fi. Now we can finally make that happen.One of the reasons why I started the company is to actually go build this future rather than just talking about it” he said.

Pat O’Donnell is a seventh-generation Western Pennsylvanian and his company Shelfmark(opens in new window) is committed to staying here — just like he did after graduating from CMU’s Tepper School of Business(opens in new window).

“I would say to anyone coming to CMU Startup Week, bring your friends, check out the city — we are a tight-knit community who lift each other up. There’s talent here, a pipeline. And to recent graduates, stay here. There’s never been a better time to stay in Pittsburgh,” recalled O’Donnell, CEO of Shelfmark.

At CMU Startup Week last year, he was encouraged by the vibrancy and creativity all around him.

“It was so great to meet new founders at their early stages and how they were approaching building their companies. Their ideas and enthusiasm rubbed off on me,” O’Donnell said.

Shelfmark is a seed-stage startup grounded by Carnegie Mellon roots.

“My co-founder, Craig Markovitz, is a professor at Carnegie Mellon,” O’Donnell said. “I met some of our first employees and earliest collaborators all at CMU over the years. And so we've made a point to stay super close to the Carnegie Mellon ecosystem since the beginning of Shelfmark. And of course, that includes attending Startup Week.”

For O'Donnell, CMU Startup Week rekindled the excitement and spontaneity that often define the earliest days of a startup.

“As a 4-year-old startup, sometimes, you know, you miss that energy of the early days of getting started, of working on just a simple idea. And so for me, every time I go back to Carnegie Mellon, I feel like I'm refreshed with the energy of folks starting from the ground up. And I think that's especially true at Startup Week,” O’Donnell said.

Since attending, his company has doubled in team members and revenue. They’ve added Fortune 500 customers, and the program is installed in businesses all over the world. Through a state-of-the-art AI visual inspection program, Shelfmark solves production-line problems in manufacturing right at the problem point so that lines run smoothly, freeing up humans to do more intricate work.

O’Donnell sees manufacturing as one of many industries that could actually add jobs because of AI. “We have clients with a 40% rate of open roles because AI has freed up new roles and created new opportunities,” he said.

As an industrial engineer, he says the Tepper School taught him things about entrepreneurship that he’d not thought of before. “That’s the thing about CMU. You learn to solve problems, not just invent something to invent something. We went out into the real world and asked people in industry what they needed. Manufacturing is a natural for our sensibilities in this region, and so we were able to put into practice what we heard from customers,” he said. “Tepper taught me so much about entrepreneurship that really applied to helping existing businesses.”

“Being a founder can be lonely,” he said. “It’s great to be around other people going through the same thing, and to support each other.”

The communal feeling of the City of Pittsburgh brimmed at CMU Startup Week, O'Donnell said.