Our Founding Story: How Uneven Treatment Patterns & Disease Definitions Inspired NetraMark’s Founding

In this post, you’ll find:

• The Idea that Inspired NetraMark
• 10 Years of Development
• NetraMark’s Mission Today and Vision for Tomorrow

Netra नेत्र is derived from Sanskrit, meaning “eyes” and symbolizes vision, insight, and perception.

The Idea that Inspired NetraMark

In 2009, Dr. Joseph Geraci was helping companies develop algorithms for drug discovery and advising pharmaceutical firms on how to better understand their patient populations. Through this work, Dr Geraci realized: poorly defined patient populations were at the heart of the clinical trial failure crisis. The problem wasn’t simply algorithmic—it was structural. He began exploring novel mathematical paradigms that could augment both human and machine intelligence across domains such as genetics, epigenetics, transcriptomics, the microbiome, imaging, and clinical records. 

The spark that led to the creation of NetraMark occurred during an analysis of an ovarian cancer dataset in which only half of the patients had survived treatment. Dr. Geraci was struck not just by the outcome but by the overwhelming complexity and heterogeneity of the patient profiles. The prevailing definitions of disease were obscuring meaningful patterns—misleading machine learning models and leading to overfitting. This wasn’t limited to oncology; similar failures were apparent in neurodegenerative and psychiatric datasets. The root problem he identified was that the taxonomy of disease itself was introducing flawed class structures for supervised learning, particularly harmful in small-sample clinical datasets. 

Throughout his career, Dr. Geraci observed this same pattern of failure as a recurring challenge playing out across the industry, especially in drug development clinical trials. When a drug failed to demonstrate separation over placebo in a trial, it was often shelved—despite evidence that it did work for a specific, often hidden, subset of patients. This pattern resulted in devastating realities where treatments with real potential were lost, and patients who might have benefited were left with no options. This led Dr. Geraci to propose a bold hypothesis:

If one could develop a mathematically-augmented artificial intelligence capable of reorganizing patient datasets in a way that would empower advanced machine learning techniques to understand heterogeneity and overcome the oversimplified, blocky labeling schemes that were misleading entire research programs, we could uncover the true advantage of a drug. With that insight, we can clearly define the biologically meaningful patient subpopulations most likely to respond—and use that knowledge to enrich and rescue failed clinical trials.

A significant share of late-stage failures stemmed not from ineffective drugs, but from poor patient stratification—treatments that worked in subgroups failed to show benefit in the full, misclassified population.

10 Years of Development

While AI in healthcare was evolving rapidly, no AI solution existed that could reliably detect response patterns in the small, complex, and noisy heterogeneous datasets that characterize real-world clinical research and predict which patients would actually respond to treatment. Recognizing this as a critical gap, Dr. Geraci began to develop the technology to support his hypothesis.

Many companies focused on large-scale automation—some aimed to accelerate drug discovery by ingesting massive datasets, while others applied AI to optimize operations like patient recruitment through electronic data capture (EDC) systems. 

Dr. Geraci began to develop NetraAI, coding the early prototypes himself in the first year and a half. Fueled by the clear and urgent insight that clinical trial failure rates remained stubbornly high, with over 85% of trials failing to reach approval, he was able to raise a seed round of funding and assemble a team.

Over the next decade, meticulous work contributed to what is now NetraAI, a mathematically augmented AI system designed to discover explainable subpopulations within small datasets.

Rather than relying on brute-force searches through all possible variable combinations or settling for shallow correlations, NetraAI identifies combinations that interact in subtle, often non-obvious ways to shape a patient’s response to treatment. At the heart of the system is a long-range memory mechanism that allows it to remember and reinforce meaningful interactions across the data landscape, even when they are separated by noisy or sparse dimensions. It’s unique learning architecture suppresses spurious models and guards against overfitting by continuously penalizing explanations that fail to generalize across evolving subpopulations. This breakthrough framework cuts through clinical data complexity to uncover the latent structure of disease, enabling sponsors to design smarter trials with a clearer understanding of which patients are most likely to benefit.

NetraMark’s Mission Today and Vision for Tomorrow

After years of validation against real-world data, including failed Phase 3 trials, NetraAI proved its ability to uncover hidden subpopulations that would have benefited from the experimental therapy.  What began as a scientific hypothesis evolved into a full-fledged company. After more than a decade of dedication and funding backed by early partnerships, NetraAI was ready to impact the pharmaceutical industry in a revolutionary way. 

Today, NetraMark has grown into a proven AI solution provider used by leading pharma and biotechs in the pharmaceutical industry, looking to learn more about heterogeneous disorders to enhance how we battle disease and produce more positive outcomes. NetraMark enables trial sponsors to de-risk their studies by identifying patients most likely to respond—before costly recruitment and enrollment begin. By doing so, it not only increases the likelihood of trial success but also supports the broader mission of introducing personalized treatment options: getting the right treatments to the right patients. 

By helping researchers better understand drug effects across complex, diverse populations today, we are changing how treatment decisions are guided for tomorrow: not by crude averages, but by deep insight into the true structure of disease.

To learn more about Dr. Joseph Geraci before founding NetraMark, check out this featured article written by his alma mater.

Featured:
The University of Toronto shared the story of where Dr. Joe Geraci got his educational start. Read how his education gave him the tools to get to where he is today:
https://www.artsci.utoronto.ca/news/accelerating-development-lifesaving-drugs-alum-joseph-geraci