Our Story

Our founders, Brett and Andrew, met while studying data science at Columbia University. Because they had both watched their loved ones suffer illness, they decided they wanted to use the data science tools they had acquired to have a broader impact on the health of humanity.

They are building EVQLV to find better and faster ways to discover, develop and deliver medicine. EVQLV’s mission is to accelerate the speed at which healing reaches those in need, whether it is those who are directly affected by disease or the loved ones who surround them.

Our Technology

We use billions of public and proprietary data points across a wide range of exponentially growing datasets including sequence, structure, biophysical properties, and more. We integrate all that data, and then we apply our algorithms and computational models in order to extract knowledge from the data.


We take an amino acid sequence as an input to our algorithms. To those sequences, we first apply our AI-generated model of evolution. The model uses large amounts of sequence data to learn the evolutionary patterns of protein formation against a target or antigen of interest. This allows us to generate 1 – 2 million evolution-informed candidates in a process which takes us less than 6 hours.

Once the candidates have been generated, they are characterized along a host of over 200 factors. For all intents and purposes, we are performing computer versions of laboratory assays, but at a scale and speed that would be incredibly challenging in a wetlab. These assays limit challenges with developability, instability, poor expression, and other downstream risks.

Because of the massive amounts of computation required to accurately predict binding and affinity for biologics, we have dedicated significant development efforts to be able to measure this key characteristic. This was accomplished by architecting the utilization of hundreds of cloud computing instances at once , which allows us to rapidly predict relative binding for hundreds of biologics at a time.

With our potential candidates fully characterized, we cluster them to produce a highly diverse and representative panel. In less than a week, our technology enables us to generate anywhere from a handful to thousands of diverse candidates. The candidates possess both genetic and biophysical diversity while retaining target specificity, all with a lower likelihood of downstream risk.

Artificial Intelligence Advantage

We use Artificial Intelligence to optimize antibodies for affinity, solubility
cross-reactivity, manufacturability, immunogenicity, specificity & stability

Accelerate the design of high quality and diverse antibodies

Expand IP portfolios by generating novel variants of existing antibodies

Quickly and efficiently improve upon existing antibodies via computational optimization

Pin It on Pinterest