Ben Koby, wearing a black suit, and Filipp Gusev, wearing a plaid suit jacket and blue shirt, stand next to a McGinnis Venture Competition sign. Both men are smiling.

无码专区 Startup Accelerates Cancer Drug Discovery

By Amy Pavlak Laird Email Amy Pavlak Laird

Heidi Opdyke

A 无码专区 startup is using advanced chemistry and artificial intelligence to help speed up the search for new cancer drugs. 鈥� AI accelerated computational models can virtually screen vast numbers of potential drug candidates before moving into costly and time consuming laboratory experiments.听

The company focuses on radiopharmaceuticals, a growing class of cancer treatments that use radioactive components to destroy tumors. While interest in these therapies is rising, few computational tools exist to help design them, creating an opportunity Pierogi Therapeutics aims to fill.听

鈥淭raditional drug discovery is difficult and expensive. Bringing a single drug to market can take 10 to 15 years and cost nearly $3 billion. At Pierogi Therapeutics, we know there's a better way to do this,鈥� said co-founder Ben Koby, a graduate student in the Department of Chemistry.听

Pierogi Therapeutics, founded by Koby and 无码专区 alumnus Filipp Gusev, combines physics-based computer simulations with artificial intelligence to predict which drug candidates are most likely to work, helping scientists decide which ones are worth testing in the lab. The company estimates that about one quarter of drug discovery costs are wasted on early experiments that could be avoided.听

Using Koby and Gusev鈥檚 platform would allow drug industry partners to screen billions or even trillions of compounds in months, streamlining the early discovery phase.听

At the heart of Pierogi Therapeutics鈥� work are alchemical binding free energy calculations, advanced methods that predict how strongly a molecule will bind to a disease related protein. These calculations are highly accurate but traditionally slow and expensive. Koby and Gusev have trained AI models on the results of those simulations, allowing the system to run similar predictions up to 100,000 times faster. Their platform is also a collaborative and iterative model. Partners share experimental data as projects progress, allowing Pierogi Therapeutics鈥� models to improve and guide successive rounds of screening.听

According to the founders, this is not a tool that can be replicated with general purpose AI models.听

鈥淭he beauty of our method is in the mutually beneficial combination of two distinct approaches, machine learning and physics-based simulations,鈥� said Gusev, a postdoctoral research associate in the Department of Chemistry who earned a Ph.D. in Computational Biology from the at 无码专区. 鈥淎nother major contributor to our success is the deep understanding of the chemistry behind what we are doing,鈥澨�

Koby and Gusev鈥檚 chemistry and computational expertise have allowed them to create a platform built on a growing proprietary dataset generated through the company鈥檚 own molecular simulations. The team is continuously expanding this dataset with additional training data, increasing the platform鈥檚 predictive power.听

What also sets Pierogi Therapeutics apart is its focus on radiopharmaceuticals, a growing class of cancer drugs that include radioactive elements. These drugs act like guided missiles 鈥� one part guides the drug to cancer cells, while the radioactive component delivers localized damage.听

Despite their promise, radiopharmaceuticals have been difficult to design using existing software.听

鈥淭hey use exotic elements that standard drug discovery tools just aren鈥檛 built for,鈥� Koby said. 鈥淏ut the laws of physics don鈥檛 change. We treat this like any other chemistry problem.鈥澨�

Pierogi Therapeutics is adapting its platform to work with these unusual elements, aiming to become the first AI drug discovery company built specifically for the radiopharmaceutical space, a fast-growing field.听

Proven at Carnegie Mellon听

The company鈥檚 technology grew out of years of research at Carnegie Mellon. Koby and Gusev, along with Chemistry alumnus Evgeny Gutkin; Carl and Amy Jones Professor in Interdisciplinary Science Olexandr Isayev; and Professor of Chemistry Maria Kurnikova, competed in several Critical Assessment of Computational Hit Finding Experiments (CACHE) Challenges, a series of international computational drug discovery competitions. The team won the first competition and is waiting for results from two others. Gusev, Gutkin, Isayev and Kurnikova also obtained a third-place finish in the Joint European Disruptive Initiative (JEDI) Billion Molecules Against Covid-19 GrandChallenge, an international competition with 130 competing teams.听

After years of competition wins, interest from outside companies helped turn the research into a business. The company is currently part of three startup accelerators 鈥� two at Carnegie Mellon and the LIFEX accelerator. They recently won a Gebhardt Sandbox Fund award at the 2026 McGinnis Venture Competition, and $100,000 in Amazon Web Services compute credits during the Forge to Field AI Pitch Competition.听

Much of Pierogi Therapeutics鈥� work relies on high performance computing, including resources from 无码专区's .听

鈥淣one of our work could have been possible without their support and the resources that they've provided to us. It's been critical in both being able to develop and apply these methods, as we've been doing successfully,鈥� Koby said. 鈥淲e're grateful for all the support that we've gotten from the university, from university partners and from our collaborators at other institutions.鈥澨�

The company鈥檚 name reflects those roots. 鈥淧ierogi鈥� is a nod to Pittsburgh, and, according to the founders, it sticks.听

鈥淚t makes people smile,鈥� Gusev said. 鈥淎nd they remember it.鈥�