Q&A with Former Googler Natasha Mohanty

We recently chatted with Natasha Mohanty, who led engineering teams on Google Search, content recommendations and personalization at Google News and Google+, on her new role at startup FEM. The video discovery/recommendation service provider that Mohanty co-founded aims to contextualize video recommendations based on what the user is doing in that moment. What have you learned from Google? One of my main learnings at Google was to think about how we would scale right from the moment we started building our product. As a result, our infrastructure is able to rapidly process hundreds of thousands of videos as well as large volumes of user behavior data, which enables us to keep our recommendations fresh and engaging. We have also optimized our system to serve millions of video recommendations at low latency. This is essential as users increasingly consume videos on mobile devices. During my time at Google News, I learned the value of recommendations that would surprise and delight the user, helping users discover stories that resonated with them. Video consumption is fundamentally a lean-back experience where users want to magically discover content that they would enjoy without having to put in the effort to search for it. Our algorithms understand how videos can appeal to users almost at an emotional level, which leads to a deeply engaging experience. What is the biggest challenge now in terms of video recommendation and personalization, whether it’s linear or over-the-top video services? Video creation and consumption has been exploding in growth. However, the space is very fragmented with a number of different players in the ecosystem, which makes it hard for users to enjoy a seamless personalized video experience across their desktop, tablet, phone and TV. As video formats get standardized, it will become a build experience that travels with the user wherever they go. In an ideal scenario, I want to receive notifications on what to watch based on my complete video consumption behavior, but tailored to the device that I am using right now. Our algorithms currently adapt to the time of day, context and device the user is on to provide the most relevant recommendations at the right moment. We are taking this a step further by enabling users to build a playlist of videos to watch from across the web spanning multiple platforms. What will you be focusing on at FEM? Our core DNA is our proprietary video recommendations and personalization algorithms to surprise and delight the user. Our secret sauce has enabled us to drive 5x views and repeat engagement with videos for our partners. The algorithms go beyond traditional keyword approaches to map users’ underlying motivations with content they would enjoy. Brands care deeply about these axes and our approach allows us to maximize the impact of branded content. Our platform provides a seamless video experience across multiple platforms and can be embedded onto a publisher site with less than 10 mins of work. We have invested a lot into making the integration easy and lightweight, including automatically learning a site’s UI so that our platform looks and feels exactly like the site’s content.