Ashish Rastogi is a Senior Data Scientist at Netflix in Beverly Hills, California, where he works on applying machine learning methods to value content and optimize Netflix's global catalog. Prior to Netflix, he held various positions at Google, Goldman Sachs and most recently Bloomberg, focusing on the application of large scale statistical inference methods to a variety of domains, including web search and financial markets.
We sat down with him ahead of their presentations at the Predictive Analytics Innovation Summit, taking place in San Diego this February 22–23.
How did you get started in your career?
While I have been generally interested in Computer Science & Programming since high school, my foray into data science and machine learning had an interested trajectory. I studied machine learning through grad school. During the course of my Ph.D. I was working on various problems in theoretical machine learning, spending my summers at Google's research lab in New York. When I graduated from NYU in 2008, machine learning wasn't a thing, at least not as big a thing as it is today. When I was studying ML, I had no idea that it would become so important!
What trends should we watch out for in 2017?
The trend in machine learning has been strongly in favor of neural networks. If 2016 was the year when Alpha-Go bested world's best Go players, I think 2017 will be the year where we make continued progress towards neural-networks based technologies entering the every-day world. I'm thinking better personal assistants (Amazon's Alexa and Google Home are already great successes), greater progress towards autonomous vehicles, machine learning in health-care, etc.
Can analytics be fully automated?
People way smarter than myself have been thinking about this. I really like Andrew Ng's perspective: 'if a typical person can do a mental task with less than one second of thought, we can probably automate it using AI either now or in the near future.' A lot of analytics that are done require way more than one second of thought! There is all sorts of 'business knowledge' that gets applied in how to transform data before it can be piped through a data science framework. The 'intuition' that goes into those processes will likely not be automated.
What are the dangers of machine-based innovation?
There are a lot of open questions when it comes to machine-based innovation. Explaining the predictions of machine learning models remains an area of active research. Ensuring that machine learning models are 'fair', in that they do not use predictors that use (or are correlated with) protected categories is extremely important, and will likely have a massive impact in how these techniques permeate the insurance and healthcare industries to name just a few. I'm personally very interested to see how we trade-off model accuracy with these social concerns of using the models in practice.
What will you be discussing in your presentation?
In my presentation, I will be giving an overview of how Netflix uses data science to make content decisions.
You can hear more from industry leading experts like Ashish at the Predictive Analytics Innovation Summit. View the full agenda here.
BONUS: Watch this presentation from last year's Predictive Analytics Innovation Summit by Alberto Rey-Villaverde, Head of Data Science, EasyJet