Espresso AI is the first compute optimization system built from the ground-up to leverage the code analysis and code generation abilities of large language models. Our state-of-the-art models enable us to automate the kinds of optimizations that, until now, were possible only with massive investments of time from the top handful of performance engineers in the world.
Today, we are excited to announce that we have raised over $11 million in funding, including a seed round led by Daniel Gross and Nat Friedman, and a pre-seed round led by Matt Turck at FirstMark, with participation from industry leaders including Tasso Argyros, Spencer Kimball, Calvin French-Owen, Tristan Handy, Aston Motes, and many others.
Our First Product: 5x Cheaper Data Warehousing
Our first product is a data warehouse optimization platform, currently available for Snowflake. We are leveraging our models’ understanding of SQL to optimize individual queries on the frontend and, at the same time, optimally allocate physical compute on the backend. The result is an instant, significant reduction in Snowflake operating costs.
Our MVP can be set up in ten minutes and is reducing costs for our customers by up to 80%. If Snowflake is a major cost for your organization and you’d like to dramatically increase your data warehousing efficiency, please contact us at: snowflake@espresso.ai.
Our Vision: 1000x Faster Compute
Large language models are a fundamental breakthrough in AI. For the first time, models can deeply understand code in the same way that human programmers do. We believe that neural optimizers built on this understanding will leapfrog legacy optimization tools and usher in a paradigm shift in computation.
Models like AlphaGo and AlphaFold demonstrate the potential of machine learning to reach superhuman levels of performance on tasks at the edge of human intelligence. Our research program aims to build such a model for compute: one capable of superhuman code optimization.
We expect our work to drive an increase in software performance not seen since the end of Moore’s law, and to result in a 1000x acceleration of compute.
Our Team
We are a small team of elite AI and performance engineers from Google, Apple, and MIT. Our founders have worked on natural language processing in Google Search, core systems performance engineering in Google Cloud, and early code-LLM research in Google Deepmind.
We are hiring world-class engineers to help us build the future of compute. If you’re interested in joining us, you can learn more and apply here: https://espresso.ai/careers.
Frequently Asked Questions
Once you sign up, you'll receive a savings estimate telling you in advance how much you can expect to see in savings.
Our models are continuously calibrated with production Snowflake data to ensure our savings numbers are accurate.
The best way for you to judge accuracy is to compare our upfront savings estimate to the savings you see in production when we first turn on.
We also encourage users to run A/B tests once we've been on for a few months: shut Espresso off for a week and see how your actual spend compares to our savings calculation.
Our models calculate savings by analyzing your workloads one query at a time. When your team makes optimizations we'll be able to see it in the logs: you'll be running fewer queries, faster queries, or using smaller warehouses. That enables us to accurately understand which savings came from our optimizations and which savings came from yours, and to charge accordingly.
Yes, we offer a money-back guarantee period. This lets you see savings live in production before committing to pay.
Never miss an update
Subscribe to our newsletter. Get exclusive insights delivered straight to your inbox.