Welcome, Ensemble!
Democratizing state-of-the-art machine learning.
- Founders: Alex Reneau, Zach Albertson
- Sector: Artificial Intelligence / Machine Learning
- Location: San Francisco, CA
The Opportunity
The past 24 months have produced a cycle of rapid technological advancements in artificial intelligence (AI) and machine learning (ML) that continue to push the boundaries of what’s possible. The AI age has led to the democratization of state-of-the-art models and modeling techniques that stand to make building with AI and ML easier than ever. As a result, enterprises can no longer treat the deployment of performant and accurate models as a competitive advantage but as table stakes to remain competitive in the market.
Unfortunately, taking models off the shelf and pointing them at jobs to be done hasn’t proven to be easy. If organizations desire an accurate, impactful, and predictive model output, not only do they need to bring their own datasets, but these datasets need to be robust, complete, and high-quality enough to drive insights. When this isn’t the case, practitioners spend more time filling in data gaps or developing increasingly sophisticated models to compensate, both of which are time-consuming and expensive.
The Solution
Ensemble allows organizations to achieve more with less data. Their team has developed a net-new representation learning algorithm that learns to approximate hidden relationships in data by outputting embeddings, effectively distilling signal from noise to allow ML models to achieve state-of-the-art performance — even with limited, sparse, and high dimensional datasets.
Similar to other deep learning approaches, Ensemble’s algorithm can account for complex, non-linear relationships. But unlike other deep learning approaches, Ensemble does this through a lightweight data transformation. By finding a way to distill a ton of complexity into a data representation instead of using a model, Ensemble enables data scientists and MLEs to build great models even in hard-to-model problem settings.
This frees up data scientists to focus on experimentation and further advancing their existing pipelines. It also makes ML viable for previously un-modelable problems and unlocks new modeling capabilities.
Why We’re Backing Ensemble
We met Ensemble co-founders Alex Reneau and Zach Albertson at a time when the default solution to bad data was more data. We were struck by their willingness to depart from the status quo of generative AI native tooling and methodologies to build a product rooted in traditional machine learning and data science principles.
This opinionated approach stems from Alex’s time in the research lab during his Ph.D. program at Northwestern, where he and his team published several notable papers defining programmable feature engineering for time series modeling. Alex’s findings were reinforced by what Zach saw firsthand when implementing AI and ML for his customers as a consultant. Together, they have the technical and practical expertise to deliver their frontier research across a broad swath of industries and use cases. Perhaps most importantly, Zach and Alex are incredibly humble and maniacally focused on making an impact for their customers.
What’s Ahead?
Ensemble’s proprietary technology and approach to data viability allow enterprises to solve old problems better and new problems for the first time. As organizations harness the power of AI and time-series data for high-value predictive workflows, Ensemble is incredibly well-positioned. Over the past 12 months, we’ve watched additional industry focus and resources pour into this area, including from Salesforce’s very own AI research team, who published Moirai this past spring —another validation point for what Ensemble has set in motion.
We look forward to partnering with Alex and Zach for years to come and were honored to invest alongside our friends at M13, Motivate, and Amplo.
Welcome to Salesforce Ventures, Ensemble!
Related Articles
