Tom de Godoy and Jeremy Achin are data scientists turned entrepreneurs. Both individuals met back in college when attending the University of Massachusetts nearly two decades ago. Being adequately-versed in machine learning models and data science, their move to establish DataRobot came out naturally.
Jeremy and Tom co-established DataRobot in 2012 with the former taking up the chief executive officer role while the latter assumed the position of chief technology officer. In Kaggle, the duo has been ranked as high as the 20th place, which is no mean feat considering that the community boasts more than 800, 000 registered data scientists.
DataRobot has been recognized among CB-insights’ leading 100 AI companies. Also, the startup has more than 400 strong employees globally, half of whom are established data scientists. Tom revealed that the motivation to start DataRobot was triggered by a Kaggle data science contest.
Even though the duo had only two weeks to develop a solution, they managed to clinch the sixth position in the competition. This accomplishment catapulted them to join additional competitions in several verticals, leading them to get three consecutive wins.
“Jeremy and I were competing together as a team around 2011. It was during the competition that we had the idea for DataRobot because it was clear to us then that every business was generating a lot of data,” said Tom. He added, “But very few companies were taking advantage of the data with predictive analytics because it requires very highly trained data scientists to do it.”
“Since the demand was not being met by the supply of data scientist, we saw an opportunity to create a platform to leverage AI (artificial intelligence) to automate this type of analysis.”
AI was Once Viewed as Science Fiction
Six years back, the idea of artificial intelligence was a new concept to many. As such, Tom and Jeremy found it difficult to pitch their groundbreaking solution to prospective investors for financial backing. “ AI was considered as a science fiction and the idea of automating machine learning was very hard to believe in for many people,” claimed Tom.
Despite the difficult experience in attracting investors back then, the current situation is different and more people are starting to believe in the AI vision. “It’s really about finding the right match. The key investors in DataRobot today are all people who bought our vision, and they know enough about machine learning and data to understand that what we were proposing was really revolutionary,” said Tom.
Back in April 2013, DataRobot managed to raise a staggering $ 3.3 million in a seed round thanks to Atlas Venture. Since that time, DataRobot has secured capital amounting to $124 million to date.
Crunching Data within Hours
Despite the current shortage of data scientists globally, DataRobot provides an enterprise-focused automated machine learning platform that allows users to rapidly create and deploy their predictive models, regardless of their skill levels. Tom emphasized that even users without any deep domain knowledge can also benefit from the platform.
Users can easily create predictive models within a short duration by taking advantage of the drag-and-drop, point-click guided menu. “Our automated machine learning platform allows enterprises to use open source algorithms, but it also automates a lot of the tasks involved in building the models, evaluating them, and then implementing it in their corporations.”
Tom also said that “ basically, it turns what used to be a project that would take months and reduce it to only two days, or sometimes even down to hours thanks to the amount of automation in the platform.”
According to Tom, a few days of training guarantees anyone the ability to carry out data analysis without external help when using DataRobot. The DataRobot university based in Singapore and Boston has trained more than 6,000 individuals in practical data science education by leveraging the automated machine learning platform.
Essentially, DataRobot’s solution enables companies to quickly find hidden data patterns, key insights and make fast predictions.
Currently, DataRobot’s technology has been utilized globally in assisting companies across multiple verticals such as financial institutions to forecast loan defaults, enterprises to set competitive pricing, insurance companies to detect fraudulent claims and many others.
Tom said that DataRobot boasts numerous financial use cases drawn from underwriting, insurance pricing, and credit risk.
Recently, DataRobot worked with one of the largest insurance providers in Singapore, NTUC Income, to assist them not only with price analysis but also enhance their business performance. Since customers have a tendency of selecting their insurer based on price points, it is important for them to create accurate commercial and technical prices.
NTUC Income can now easily evaluate whether the risk they are going to take up is priced appropriately through automating the building process and comparison of models that explore cost vs. risk.
Xavier Conort, DataRobot’s chief data scientist, was recently ranked first in the entire globe for two years in a row. “We started our Singapore office in 2013 because of Xavier actually. We made it a point to make him the first hire in the company, which turned out to be a huge manpower boost for the company. From there, we slowly worked on growing the Singapore team,” said Tom.
Back in November 2017, DataRobot invested $11.1 million to establish its Singapore-based regional headquarters as well as its research and development center. What ’s more, DataRobot prides itself on being the first few foreign companies with a regional headquarter in Singapore that is accredited under Accreditation@SG Digital programme.
This programme is known for accrediting innovative technology companies in a bid to establish credentials and position them as eligible contenders not only to large enterprise buyers but also the government.
DataRobot is convinced that such accreditation will greatly boost the company’s potential to work with the Singapore government without the need to spend a considerable amount of time trying to cut through the bureaucracy. As such, securing government projects will be much easier and faster.