Pranshul Sardana’s technology accelerates data annotation, making it easier to prepare large, complex image and video datasets for machine‑learning models to enable higher accuracy predictions. The technology will primarily target medical and pharmaceutical AI applications, where freeing up the domain expert’s time could help develop new therapies more quickly and efficiently.
When Sardana first began the Purdue Innovates Commercialization Fellowship, he had a simple hypothesis: if his team was experiencing challenges with data annotation, others were likely facing the same problem. Through the customer discovery process, that assumption proved true. It also revealed important nuances, allowing the team to discard assumptions that didn’t translate from the lab into real world use.
Since beginning the fellowship, Sardana has made significant strides to move the data annotation technology to market, including filing a non-provisional patent, applying for NSF grant, and receiving a technology refinement grant from Purdue Innovates. These milestones have provided both validation and resources as the team continues to advance the technology.
Faculty Advisors: Arezoo Ardekani and Steve Wereley
Through conversations with industry and academia facilitated through Purdue Innovates programming, we were able to narrow in on the real need and eliminate some of the assumptions we initially had.
While the underlying technology is rooted in cutting-edge research, Sardana says the commercialization journey has taught him that technical performance alone is not enough. Now focused on understanding how the technology can fit into existing workflows, the team is working to create a product that integrates naturally into the systems customers already use.
With an existing MVP ready to test, Sardana hopes to mature the technology package and integration over the next three months. As the technology continues to advance, customer feedback remains the most important ingredient in building a solution that addresses a genuine need.
Through the fellowship we’ve realized we’re not just building technology. We’re figuring out what the product needs to look like so it can work seamlessly for the people who use it. Our goal is to solve a real problem for someone else, not just demonstrate what’s technically possible.
Currently, the team is onboarding a select group of early design partners. If your team is facing annotation bottlenecks in scientific and medical imaging pipelines, such as histology, Sardana and his team are offering live demos with real datasets.
Visit the Purdue Innovates website to learn more about the Commercialization Fellowship.
Media contact: Polly Barks, phbarks@prf.org
