How HAL 9000 Altered the Course of History and My Career
Kevin Doyle, October 2020
In 2012, Thomas H. Davenport and D.J. Patil published an article in the Harvard Business Review entitled Data Scientist: The Sexiest Job of the 21st Century. This was it. My time had come. I was seven years into my data science career, scoping, building, and deploying models across retail, health insurance, banking, and other industries. I was a nerd, and I knew it, but now nerdy was the new sexy.
Fast forward three years to 2015, when the cult blog KDnuggests.com published survey results that took the wind out of my sails. 51% of respondents expected data science tasks to be automated within ten years. I felt an eerie sense of dread wash over me. Perhaps this was how statisticians of old felt when their departments first purchased an IBM mechanical punched-card tabulator (aka a computer). I can hear their pushback: “But I enjoy solving complex mathematical equations over the course of ill-defined periods by hand!” I wondered what those statisticians were doing now.
What was I going to do? Did this mean my GQ Man of the Year invitation wasn’t in the mail? I had spent years cleansing, reconciling, and wrangling data, and building, testing, and rebuilding models. Surely my careful feature engineering, carving out holdout periods, and checking for monotonicity were inextricably linked to human judgment. Surely no computer could automate my infallible* outputs. Right, HAL?
(* completely fallible)
Fast forward to 2020. So, what’s happened? First, machine learning is table stakes. Second, as businesses strive for scalability and first mover advantage, AI and automated machine learning offer a competitive edge. Forrester released an independent report noting that teams using DataRobot save $1.2 million over three years in efficiency alone. Finally, from a professional perspective, I still have meaning and direction in my career. Spoiler alert: my role now has an official title. Here at DataRobot, we call it AI Success.
For an ex-data scientist like myself, what are those intangible qualities and behaviors that enable me to lead my customers to AI Success? I boil them down into three areas:
Throughout my career, I consistently found myself bridging the gap between deep analytics, senior leadership and various other business stakeholders. This grew organically based on my technical background, clear communication, and handy management skills. My career coach shut me down when I described myself as a generalist, a jack of all trades. He disliked the negative connotation associated with the second part of the adage, master of none. For this reason, he relabeled me a multi-specialist, because I knew enough about the commercial, technical, and operational aspects to own, drive, and coordinate these projects. And it is this unique multi-skill set that places AI Success at the heart of AI solution delivery.
Strategic consulting across businesses and verticals can be challenging without the deep expertise that comes from years in a specific industry. But I’ve found two things that enable me to partner and add value to any business: an entrepreneurial mindset and logical thinking.
By approaching any engagement with enthusiasm, a passion for problem solving, and business savvy, you can logically isolate business requirements, align on value expectations, and plan around commercial and technical challenges. Through blood, sweat, and planning, you can successfully integrate AI into your business strategy.
Another professional coach left me with a key lesson in people management. Customers want any number of three things: to make money, to save money, or to look good. If you can identify what motivates your customer and nail those requirements, you’re on the path to becoming a coveted trusted advisor.
A second lesson this coach imprinted on me was that people will (mostly) only want to work with people they like. People are community-minded, and relationships are paramount to a well-functioning society. I keep these two principles in mind in every one of my engagements. Stellar customer engagement is about working together to achieve identified goals and having fun along the way.
My career journey continues, from its roots in data and statistical modeling, to value delivery and relationship building. I no longer need to spend hours in front of EMBLEM, hand-smoothing curves and choosing between a four-week lagged mean vs. five-week lagged median. I now get to focus on the stuff that really matters: making sure machine learning and AI are focused on delivering value and enabling my customers to do so. This is not the end of the road for me and many other success managers. It’s the continued progression of a smarter way of working. Thanks, Hal.