RealityEngines.AI Becomes Abacus.AI And Lifts $13M Series A Fund

by StartupWorld Staff         

San Francisco-based fintech startup Abacus.AI, previously known as RealityEngines.AI, has launched its autonomous AI commercial service into general availability. Alongside this, the startup has also bagged $13 million Series A investment led by Index Venture. As per the agreement, Index Ventures' Mike Volpi and Sherpalo Ventures founder Ram Shriram will come on-board to join the startup's board.

The US-based Minted CEO Mariam Naficy, the US-based Confluent president Erica Shultz and co-founder Neha Narkhede, California-based VC firm Sherpalo Ventures' founder Ram Shriram, the US-based Basis Set Ventures founder Xuezhao Lan, Palo Alto-based AME Cloud Ventures founding partner Jerry Yang, the US-based VC firm Decibel Partners as well as Jeannette Furstenberg also participated in the investment round. With this funding, the startup touches $18.25 million in total funding.

In 2019, Arvind Sundararajan, Bindu Reddy, Chetan Rai, and Siddartha Naidu founded AI research startup Abacus.AI. Its mission is to assist businesses in implementing modern deep learning systems into their business processes and customer experience without needing to do the heavy lifting of learning how to train models themselves. Abacus takes charge of the model training and data pipelines and for it.

RealityEngines.AI Becomes Abacus.AI And Lifts $13M Series A Fund

The startup has employed 1,200 beta testers. In the past months, the startup team majorly focused on not only helping businesses established their models but also employee them into production. Abacus.AI's prominent customers include Flex, 1-800-Flowers, DailyLook, and Prodege.

In the past recent months, the startup team also employed new unsupervised learning tools to its pre-built solutions set up to help consumers build systems for oddly detecting transaction fraud and account takeovers, for example.

The startup launched new tools for debiasing data sets, which could be employed on already trained algorithms, automatically building training sets, also with comparatively small data sets- it is one of the areas over which the startup's team has focused on its long run. The startup is now employing a few of these same techniques to handle this problem. 

In recent experiments, the startup's facial recognition algorithm enabled to highly improve its ability to detect if a Black celebrity was smiling or not. For example, although the training data set featured 22 times more, white people.


4.9/5

129 Reviews

Your rate 3 stars is recorded

About Author

StartupWorld Staff

StartupWorld is the destination to learn about new and upcoming startups making waves on the web. Being an entrepreneur is as much hard work as it is fun. No one knows it better than you do. Share your story, send it in and let us showcase it to the world.

Comments ()

Leave a Comment