In this “Lab Chat” interview we speak with Damien – new talent in the company, who’s fresh perspective provides some really curious insights into the company and industry.
Damien is from Cape Town. He has a passion for the automotive world, be it engineering or race-driving. He personally experienced the pain imposed by the COVID travel restrictions having travelled twice during the lock down (having genuinely ticked one of the series reasons for travel).
Enjoy this interview and watch out to hear more about this raising star of the DRL.
You’re quite new at Decision Lab. What made you come to work here?
Decision Lab is an industry agnostic and project-based company, and that indicated to me that there would be a great variety of work to be done. In addition, the company works with reinforcement learning and promotes research into new, cutting edge techniques, both of which excite me. It was my goal to be in practical research and development, specifically in Reinforcement Learning, and Decision Lab provides the perfect opportunity to achieve that goal.
What are your thoughts so far?
The teams were very welcoming to my arrival despite being the first consultant employed remotely. I have been able to interact with all my colleagues and chat with them all, despite only having met three of them in person to date!
I have been involved in many projects, not all of them related to my field of work and have been able to learn and even teach new techniques in the realm of machine learning. Within one week I was technical lead on a project which was exciting, not to mention a good indication of the trust Decision Lab puts in its employees, even new ones.
Is there something in your past experience that you hope to utilise at DL?
Coming from the automotive manufacturing world, I would like to utilise my industry knowledge to produce applicable RL solutions to aid in mass manufacturing. I would also like to use my knowledge of Dynamics and Control Theory more to produce AI that can directly interact with the physical world.
What do you enjoy most about your work?
I enjoy solving the complex problems Decision Lab has on offer. The great thing is that these problems vary in scope, industry and solutions required. I can be working on three projects simultaneously and each will be for a different client and require a vastly different solution.
I am also getting into things I always wanted to but never had the time to previously, such as video game development. Just through the nature of our projects I am able to learn this extremely fun skill.
Being client facing, I can directly discuss with the client their problem and provide solutions. Having been involved in several bids, I have done a fair bit of “solutioneering” which is the most fun part of being an engineer. I also get to learn more about the business aspects of Decision Lab which makes for a well rounded experience.
It’s also fantastic to work with such highly qualified and intelligent colleagues, as there is a wealth of knowledge on tap and getting support when you are stuck is merely a slack message away.
What would be your dream project to work on?
I would like to be involved in a project that directly relates to sport and motorsport in particular. I find the dynamic nature and strategic requirements of real time sport to be the perfect setting for Reinforcement Learning. As an amateur racing driver myself, I often find myself wishing for an expert strategist to plan the key move strategy wise. Strategists are expensive to employ, so an automated strategist would be a welcome addition to the amateur driver and developing this would not only help me but I think it would be fun to test and develop as well.
What is the best advice you’ve been given in your career?
“Before you can be a manager, learn how to do the technical stuff.” I believe that to accurately manage a project or business there is an intrinsic need to understand the technical capabilities and constraints of the tools on offer. In my experience, the best managers are those who have been exposed to work on the floor and know how to manage the technical risks that may arise.
And while cliché, “Variety is the spice of life” is a very important phrase. To keep the mind sharp and evolving it is better to be constantly trying new things as exploration is the mother of innovation.
Where do you see the industry in 5-10 years?
RL will become a part of practical industry and will be less of an experimental technique used only in video games. I also see new legislation covering the liability and rights of AI agents to be improved. Particularly in the defence industry, the question of who is liable in combat situations becomes a pertinent issue.
I also think that explainable AI will become one of the most researched topics in AI, if not become a hard and fast requirement for any customer. I am positive that explainability standards will be developed and these will need to be adhered to rigidly to secure future contracts. It would be great if we could help formulate those standards!
What industry or technical challenges intrigue you the most?
I enjoy dynamic environments with a high level of uncertainty. In particular I like working with a client in a new-to-me industry as it forces me to learn about the domain and then use RL to figure out the best solution. I specifically like working with environments with regions of partial observability as then we need invoke the concept of curiosity in AI agents and this is starting to get to a more “human-like” AI.
Anything else you’d like to tell us?
I am keen for life to return to some sense of normality and being able to travel again. Having lived in Europe for three years now, 2 of which have been during COVID, I haven’t been able to take advantage of the cheap travel fares and see what the continent has to offer. Hopefully, that will change soon and a holiday in Greece or Spain will be on the cards! Otherwise, I’ve taken up karting and that’s going well so hopefully it gets even better as the season continues.