Vinod Iyengar crushes silos with deep learning

Data is the future. As we generate more of it, those who can do the most with it, who see unique correlations and act on them, have an advantage.

To maximize that advantage, companies need computing power and they need data science, two things out of the reach of many businesses.

That is where steps in, data scientist and director of product marketing Vinod Iyengar said. provides deep learning services to 169 Fortune 500 companies, eight of the 12 largest banks and seven of the 10 largest insurance companies.

As enterprise artificial intelligence services improve, the typical workflow is eliminating silos and bringing all data into one place, Mr. Iyengar said.

Vinod Iyengar

Now, what does a company do with all of that data?

“You need a platform to run machine learning algorithms on big data,” Mr. Iyengar responded.’s newest product does precisely that, he added. Driverless AI lets companies with a small data science department or none at all realize the benefits interpretation of big data brings. Running up to 50 times faster than other products on the market, it automates fraud prevention, credit scoring, and product recommendations while providing understandable results for non-technical consumers.

The most effective data scientists are like detectives that use data to solve problems, Mr. Iyengar said. Looking for interesting features in data, they develop a series of models to see which one best addresses their hypothesis. That takes time to complete.

Driverless AI reduces that time dramatically, Mr. Iyengar said, by producing the ideal model in a fraction of the time. They accomplish this through an accelerated graphics processing unit (GPU) that shortens computing time.

“By using the GPU we can increase the speed of the process,” Mr. Iyengar said. “Results can come in minutes versus hours and days.” was the first company to determine machine learning would benefit from GPU deployment, Mr. Iyengar said. That allows it to effectively manage larger and larger datasets that are becoming commonplace.

It also allows companies to quickly react to changing customer demographics. Say a company that had long seen a 50-50 client gender split experienced an influx of female customers after a recent marketing campaign. With technology, they can quickly reconfigure marketing plans to ensure they effectively communicate with their new clientele.

Other suspicions may be true for a sliver of the population. The system quickly identifies which ones and goes to work to produce suitable models for the remainder. As behaviours change, such as when fraudsters adapt to new AML regulations, the system can easily run simulations with new parameters and learn as it goes.

“With machine learning, the AI engine is constantly learning not basic rules, but more complex ones,” Mr. Iyengar said. “Our rules are simple because it’s tough for humans to develop more complex ones.

“Artificial intelligence is the only way to address this.”

Notify of
Inline Feedbacks
View all comments