Interview: John Funge, CTO, Knack on Why Gaming is the Next Big Thing for Hiring
Tags: Advice, Analytics, Big Data, Gamification, Gaming, Hiring, John Funge, Knack, Skills, Startup, Tools
We discuss the gamification of hiring, founding story of Knack, applications of Predictive Human Analytics, challenges, Big Data tools and technology used, key qualities sought in data scientists, career advice and more.
John Funge is CTO and co-founder at Knack, a company that matches job seekers to jobs through analyzing big data from Knack games to discover a person's unique combination of strengths, talents, and personality traits. John was previously Director of Engineering at Netflix. Before joining Netflix, John co-founded AiLive, a real-time machine learning company that became synonymous with automated motion recognition middle ware that was used in top-selling games for the Wii and PS3.
John grew up in the UK, obtained his MS from Oxford University and his PhD from Toronto University, he has authored three books on Artificial Intelligence and computer games (Game AI), and taught at the University of California Santa Cruz where he was an Adjunct Professor.
Here is my interview with him:
Anmol Rajpurohit: Q1. How and when was the idea for "gamification of hiring" perceived by Knack founders? How has the journey been since 2012, when Knack was founded?
John Funge: The original idea came when Guy Halfteck (CEO and co-founder) started to think about how hiring could be improved after he was turned down for a
hedge fund job. Early discussions and experiments made us realize just how rich game play data is for learning about people’s personalities and traits. The journey since 2012 has been about building an amazing team, proof of concepts, and then consumer and enterprise-ready products.
AR: Q2. How do you define "Predictive Human Analytics"? Are you solely focused on hiring or also looking at other application of Predictive Human Analytics?
JF: We define Predictive Human Analytics as the ability to accurately predict personality and traits. In an employment setting, that means being able to predict job performance based on game play data. Right now we are focused on hiring, but we realize there are many other areas, such as college admission and team building that we are interested in working on in the future.
AR: Q3. How do games provide personal insight? What do you think about the possibility that the skills one displays while playing a game, one might not display the same in real-life?
JF:
There is a great quote, sometimes wrongly attributed to Plato that you can talk to someone for hours and learn almost nothing about them. But play a game with them for even a few minutes and you know all about them! Games can provide insight on all aspects of a person's traits and personality.
For example, we have used our games to infer cognitive ability, conscientiousness, leadership potential, creativity as well as predict how people would perform as surgeons, management consultants, and innovators. Game data is behavioral and so it makes intuitive sense that it tells you a lot about a person’s actual behavior in real life. Moreover, we have performed many rigorous validation studies to make sure our inferences are applicable to the real world.
AR: Q4. What are the key challenges in deriving behavioral insights from gaming analytics?
JF: Our games provide a high volume of data and so we need to rapidly distill that information and make inferences to provide a compelling customer experience where people play games and get timely feedback.
AR: Q5. What are the Big Data tools and technology that you use? How do they match your goals and day-to-day operations?
JF: For exploration we make heavy use of Matlab (and the open source version Octave). In production, we have taken advantage of many of the great tools that AWS offers for horizontal scaling in dealing with big data such as load-balanced auto-scaling groups for EC2 instances, the DynamoDB NoSQL database, and SQS. We have used Clojure running on the JVM for some data processing but are currently excited by some of the great analytics tools available for Python.
AR: Q6. What key qualities do you look for when interviewing for Data Science related positions on your team?
JF: We like to hire people who can perform great exploratory analysis but then also take those ideas and put them in production. Those people are hard to find but so far we've been fortunate to hire some outstanding data scientists, the best I’ve ever worked with in fact.
AR: Q7. What is the best advice you have got in your career?
JF: That you need to constantly strive to hire great people who “raise the bar” and not just settle for people who might be useful.
Related:
