Skip to content

About Featrix

LLMs and Generative AI have kicked off an unprecedented wave of innovation. Early applications of GenAI focused on processing previously untapped unstructured data, yet structured data still drives a majority of the world economy. Current approaches to predictive analytics on that structured data tend to be hand-crafted labors of love, which simply won't scale going forward.

Our story


That's why we started Featrix—to tackle the need for automated analytics and predictive AI for the next generation of computation. Vector-based computation opens up new possibilities, and not just for unstructured data. Our goal is to make predictive AI models on structured data as easy as LLMs have made answering questions in natural language.

As the foundation, we developed the Featrix AI SDK, which brings the power of vector embeddings to tabular data. At its heart is our Featrix Analytical Pre-trained Transformer (APT), which combined with either off-the-shelf or custom models delivers a cutting edge AI system and developer experience that requires no additional tooling for data processing or model monitoring.

Building on our SDK, we are providing solutions for predictive analytics, and we have started with lead qualification and website personalization. We take care of it, from end to end.

Our team: 100 years of B2B software experience


Mitch Haile

CEO & cofounder

Mitch has been delivering data solutions to the midmarket and enterprise for over two decades. With deep experience in compression, ETL, backup, virtualization, data monitoring, Mitch has over 36 issued patents and was among the first ten engineers at both Data Domain and SnapLogic. He later founded Pancetera (sold to Quantum). Mitch has run marketing, product management, and engineering teams.


Pawel Zimoch

CTO & cofounder

Pawel dropped out of his MIT PhD in mechanical engineering to follow his love of entropy from thermodynamics to information theory to machine learning. Pawel previously worked at Gamalon, a natural language processing AI company.


Michael Rowan

VP of Engineering

Michael has an extensive background in creating and operating businesses in distributed systems, analytics, high availability, and data protection, delivering transformative outcomes that run the gamut from startups to being a Director at Amazon.  He has 16 patents across multiple technology segments and has built dozens of high-throughput data systems that process billions of data items.  


Bernhard Suhm


Bernhard has applied AI in a variety of use cases, most recently in engineering applications at MathWorks and to semantic search at Elastic. He has led analytical teams, managed products and technical marketing. He holds several patents in multimodal interaction and call center analytics, and led the deployment of SaaS analytics to some of the largest contact centers in the US.

Company news