Aquarium

AI is well underway to becoming a ubiquitous technology for businesses of every industry, whether it’s manufacturing inspection, self-driving technology, or agriculture.

However, most companies getting started with AI struggle with two main things: 1) Identifying the right training data and tooling to teach their models; 2) Getting their models to work well in production.

There’s a common misconception that the more data you feed into a model, the better the model. In reality, there’s simply too much data to label and train on. It’s important to target data labeling towards the examples that would most improve your model, otherwise your model performance will plateau.

The secret to making machine learning work, learned through hard lessons in production deployments, is to set up a virtuous cycle of iteration where it’s easy for ML teams to find problems with their models, fix those problems with data, and then retrain + redeploy better models.

We saw this problem first-hand at Cruise and that’s why we built Aquarium. We want to take the same type of technology that FAANG / self driving giants have already built and bring them to the rest of the people in the AI industry who are trying to make their models actually work.

Milestones
  • Founded 2020
  • Partnered 2020
Team
  • Peter Gao
  • Quinn Johnson
Partner
Sorry. This page is not yet translated.
Submit
Reset
Sort
Filter
  • Consumer
  • Enterprise
  • All
  • Devices
  • Financial
  • Gaming
  • Healthcare
  • Marketplace
  • Media
  • Productivity
  • Search
  • Social
  • All
  • Apps & Services
  • Data & Analytics
  • Dev Tools
  • Hardware
  • Healthcare
  • Infrastructure
  • Marketing
  • Networking
  • Security
  • Storage
The Dentmakers.

Since 1972, we have partnered early and at every stage of growth with the founders of companies that now have an aggregate, public market value of over $3.3 trillion. Meet the companies below.

Select Categories
Filter By Partner
Viewing Companies