Mar 18, 2020

Our Investment in nextmv

Most of us have taken an Uber ride. Pressing a button on your phone and having a car appear a few minutes later may not seem like magic. But it really is. How does Uber manage to match millions of passengers with drivers - all in real-time?

Alex Iskold

The answer is by using a Decision Engine.

A decision engine is a computational platform that takes into consideration hundreds or even thousands of data points, including ratings of passengers and drivers, weather conditions, distance, and passenger preferences. The list goes on.  

Decision Engines are magical, but they are really hard to build. These engines have the ability to ingest millions of data points in real-time and make decisions that impact operational efficiency, and ultimately, the bottom line of a business.

As more and more businesses around the world move their delivery, operations, supply chains, and logistics to real-time, the need for Decision Engines grows.

This is exactly the opportunity that Carolyn Mooney and Ryan O’Neil, who previously ran the Decision Engineering team at Grubhub, recognized.

Together, they created the first real-time decision automation platform. nextmv’s cloud native modeling, optimization, and simulation tools let you encode, test, and manage repeated (and often operational) decisions easily, so your business scales with increased capacity and lower cost.

nextmv’s engine is aimed towards developers and comes with a bunch of flexible modules and components:

  • Simulation Engine: On-demand simulation engine for customer use cases, including re-usable simulation system components, actors, and models. These convert a customer system state and recent data observations into estimates of KPIs.
  • Staging Environment: Projects potential near-term system states. Tests multiple algorithms against each potential scenario and estimates individual and aggregate KPIs.
  • Decision Models: Algorithms for automated decision making in the use cases. These may be customer-specific and will be either provided by nextmv on a consulting basis or pre-exist in customer environments. A key integration point with the nextmv platform is implementing a contract between the simulation and decision models.
  • Algorithm Selection: Recommends algorithms, parameters, and objectives in real-time based on staged models, target KPIs, and configurable business rules.
  • Real-Time Sampling: Randomly samples recent data from customer event streams.

The nextmv platform has already signed on several notable customers and the team just finished Y Combinator’s W20 program.  

The team at 2048 Ventures is thrilled to partner with Carolyn and Ryan to help create the next generation of decision making infrastructure for businesses around the globe.

Subscribe to our newsletter

Signup to get quarterly updates and important announcements from 2048 Ventures.