Thank goodness for the data analysts — people who make sense of an ever-expanding, ever-changing ocean of messy data to help all of us make better decisions. Our industry has spent decades steadily building more powerful tools for this group. We do that because our society has only so many people talented enough, and such people can only accomplish so much work in a day. And yet, those people will be the first to say that for all the tools, their work is far from automated and, in fact, contains plenty of manual effort. For example, the effort to clean, normalize, prep, and transform data classically consumes the largest percentage of an analyst’s time, which they would preferably spend on producing insights.
Big silos exist in organizations between data analysts, data engineers, and data scientists. Analysts cannot produce insights at the speed they want because of a lack of powerful tools, particularly for data transformations, resulting in inefficient back-and-forth between teams. Adding “yet another tool” to this status quo has been a bit like the definition of insanity, trying the same thing and expecting a different result.
The rise of large-scale foundation models has created an opportunity to try something different. The reasoning and generalization capabilities of foundation models are categorically new. For the first time, we can empower data analysts with collaborative assistants that follow the nuanced lead of a human at the keyboard.
Given this, Madrona is thrilled to lead the Series A financing in Numbers Station. Using foundation model technology pioneered at the Stanford AI Lab, Numbers Station’s mission is to accelerate insights and data workflows by democratizing access to the modern data stack — starting with data transformation — for a non-technical audience of business and data analysts.
Numbers Station is building and applying its own foundation model technology to radically improve data analysis for all of us. Specifically, the company is building a platform for data stack automation, starting with data transformation with foundation models that will enable data analysts to transform data in the warehouse using a natural language. It will also generate SQL or Python code for transformation and provide machine learning transformations like classification or extraction into the modern data stack as part of an analyst’s workflow. While there are various basic data prep tools in the market today, we believe Numbers Station is providing a zero-to-one improvement to unlock this massive market, which is only the first step in their ambitious goals.
Equally or more important than the company taking on these major market and technological trends, we are backing Numbers Station because of the world-class team. Three co-founders received their Ph.Ds. at Stanford, working in the Stanford AI Lab under the fourth co-founder, Chris Re. CEO Chris Aberger went on to work for four years at SambaNova Systems, leading its ML software team. Chief Scientist Ines Chami was a primary author of foundational papers on which Numbers Station continues to build. Sen Wu is the chief architect. And, these founders have already built a high-performing initial team of 10x engineers and ML experts. In addition to their world-class AI pedigree and technical prowess, this team is deeply customer driven and focused on building a scalable business, not just great tech. To this end, they have been working with a stable of blue-chip financial services, software, e-commerce, and consumer companies.
The company also represents a bit of a reunion for Madrona. We were privileged to invest in the initial round of Chris Re’s first company, Lattice Data, which Apple acquired. We were also the first investor in Algorithmia, where Diego Oppenheimer was co-founder and CEO. We are joining Diego on the Numbers Station board — he represents FactoryHQ, which incubated and launched the company with the help of Factory founders Chris Re and former Lattice Data CEO Andy Jacques. We are also thrilled to have Algorithmia co-investor Rama Sekhar, from Norwest, co-invested in the round.
We are excited that Numbers Station is the latest Madrona investment at the intersection of machine learning and data, joining other portfolio companies like Runway, MotherDuck, and Snowflake. We continue to aggressively pursue opportunities where foundation models are creating new intelligent applications, as well as the enabling technology and tooling needed to unlock the power of generative AI for more businesses and users.
In closing, let’s come back to the analysts. The constraints these people face mean that our society and economy must accept fewer data-driven decisions. So, let’s fix that. The radical power of foundation models presents a chance to move the needle here in a way we have not seen for years. We can’t wait for the road ahead and the opportunity to help this team serve customers and bring their exciting vision to reality!