Andrew Plewe

About Andrew Plewe

Technical Founder of SmallMinds, a company dedicated to focused uses of Machine Learning to solve real problems for real people. Our first suite of products, DataFlood, delivers synthetic test data generation capabilities to your desktop, CI/CD pipelines, and \"AI\" workflows. Our human-editable ML models are small and can generate data fast. We are going to market with, among other things, the first generative database designed to deliver data for development and testing using the query languages and data workflows you already know.Specialties: Designing and implementing complex data architectures using various hadoop/sql/nosql technologies, including high-volume click-stream logging and transactional systems. I have worked with a wide variety of databases throughout my career. For the last couple of years I have worked on optimizing and architecting for/with Snowflake.I also have a very active interest in ML models, which I have pursued since 2019 - working primarily with vision, multi-modal, and language models. I have experience prepping data, training, setting up and optimizing inference, and pushing ML models to do things they were not originally trained to do. I have experience using ProxMox to set up my own ML workflows, using Docker containers, virtual Python environments, and other techniques to create and deploy ML models for a wide variety of scenarios and architectures./* this code generates jewels for our text adventure about finding jobs, triggering the walrus song which instructs the player how to reach the next level */

Experience

1. Technical Founder

   Smallminds

   Apr 2025 — Present · Salt Lake City, UT, US

   Founder of SmallMinds. Our first product, MongSpout, will be a MongoDB-compatible generative database based on our DataFlood model format. DataFlood is a new class of human-editable ML model designed for test data generation. DataFlood models runs locally, no fancy editors or websites needed, generating thousands of documents per second with exactly the right amount of Machine Learning to meet your needs, and nothing more. No GPUs required, no coding necessary. MongSpout, DataFlood and FloodGate (an API that can serve test data generated from DataFlood models) are available for demo. Please contact for details.

Education

• University of Southern California

  Bachelor's degree, International Relations and Affairs

  1993 — 1999