Connect with us


Chris Nicholson, Founder & CEO of Pathmind – Interview Series




Chris Nicholson is the founder of Pathmind, an AI startup that applies deep reinforcement learning to supply chain and industrial operations. Pathmind was founded to help businesses handle deep economic change and increase the resilience of their operations with AI. Chris oversees the company’s strategic vision and day-to-day execution, driving innovation and growth for Pathmind’s technology platform, and optimizing performance in warehouses and on factory floors as part of the digital transformation of business.

Previously, Chris was the CEO of Skymind, an AI company focused on deep learning. Before that, he held communications and recruiting roles for FutureAdvisor, a Sequoia-backed Y Combinator startup that was acquired by BlackRock in 2015 for $200 million. Chris also spent a decade as a journalist, reporting on technology and finance for The New York Times, Bloomberg News and Businessweek, among others.

You were previously a correspondent for The New York Times and an editor for Bloomberg News before pivoting to machine learning, could you discuss the journey behind this realignment of your career path?

When I was a journalist, in the 2000s and early 2010s, the industry was suffering as ads moved to search engines and social media, and traditional publications lost readers. I was covering finance and tech and I saw people building a lot of new and interesting things. At a certain point, I told myself I needed to get on the right side of the robots, so I moved to San Francisco.

That was 2013. There were still hacker houses in SOMA then. I moved into one of those, onto a bunk in a room with five other guys, and went to work each and every morning at a startup that had gone through Y Combinator. Working in San Francisco, I slowly got to know people and figure out what I was interested in. Back then, deep learning was still taking off. You could see the potential, but there was still a lot to build. This curiosity drove me to get involved in an AI project that eventually led to Pathmind.

Could you share some details on how Pathmind can assist businesses?

Businesses that have a physical plant to manage, such as a warehouse, a factory or a mining site, usually confront really complex optimization tasks. That is, they need to coordinate a lot of equipment to meet a goal. Many of them are trying to maximize throughput and efficiency, and minimize costs and carbon emissions. We design algorithms that can use their real data to show new ways for them to make their equipment behave. That can apply to job and machine scheduling, fleet routing and energy cost management, among other things.

What are some benefits of open source software for machine learning systems?

Software companies are all boats floating on a sea of open-source code. The infrastructure of the public clouds runs on open source. A lot of the tools that developers use to build web sites and machine learning applications is also open source. It’s a vibrant ecosystem. We are early adopters working for exciting new open source projects such as Ray, which helps you spin up distributed computing for AI workloads.

You are clearly bullish on the future of deep reinforcement learning, in your opinion what types of businesses are best positioned to take advantage of this type of machine learning?

The companies we work typically have data they’re using to track their operations and performance. They tend to have physical operations where they can control the behavior of the equipment. Our algorithms need data to understand the environment in which they operate, and they need to be presented with actions to choose from, like equipment to control, so that they can have an effect on the outcomes. Essentially, businesses that have physical operations to control, have data about those operations, and some analytics in place for that data, are in a good position to start thinking about optimization like ours.

Is there anything else that you would like to discuss about Pathmind?

One thing we do that is adjacent to controlling physical operations is to help businesses predict time series. So, data such as prices or demand will influence how a company behaves, and how much it produces. To produce the right amount, they need to know how much demand there will be. And, to be able to correctly set their prices, they have to have a read on price fluctuations. Our algorithms are able to make those forecasts.

Thank you for the great interview, readers who wish to learn more should visit Pathmind.

Antoine Tardif is a Futurist who is passionate about the future of AI and robotics. He is the CEO of, and has invested in over 50 AI & blockchain projects. He is the Co-Founder of a news website focusing on digital assets, digital securities and investing. He is a founding partner of unite.AI & a member of the Forbes Technology Council.