In this month's FACES series, we are excited to introduce Nikolaos Virtsionis, one of our dynamic Machine Learning Engineers at NET2GRID. With a robust academic background in Electrical and Computer Engineering and a master’s degree in Artificial Intelligence from the Aristotle University of Thessaloniki, Nikolaos is currently pursuing a PhD focusing on energy disaggregation. As an AI researcher, an avid guitarist, a member of NET2GRID’s company band, and a full-time dad, Nikolaos will be sharing with us his insights and learnings from the field as well as his thoughts on the future of various AI technologies.
Hi Niko and welcome to FACES #28 May! Your professional background has been informed by your studies in electrical and computer engineering and artificial intelligence. How have these disciplines complemented each other in your work?
At NET2GRID, we design and develop cutting-edge EnergyAI tools and solutions that help residential end-users reduce their energy consumption and ultimately come closer to the green energy transition. As an Analytics Engineer, I believe it is very important to possess knowledge and skills related to AI and computer science along with a good understanding of how electrical systems work. Since these problems we have to solve belong to undiscovered areas/domains, it is not always clear what piece of information is useful. Hence, every bit of knowledge could come in handy depending on the situation.
2. You are currently involved in research that focuses on using neural networks for energy disaggregation on devices with limited computational power, high-quality synthetic data generation and user insights/recommendations based on household energy consumption. Can you describe some of the key challenges and breakthroughs in this area?
All these tasks have their challenges, with one being the most crucial; the data quality. Since neural networks or any other machine learning algorithm are usually used to solve problems by example, the use of high-quality datasets, appropriate for the given task, is necessary to provide high-quality results. As a PhD researcher, there is usually a struggle to find an appropriate dataset even though there are some good public datasets for some problems (eg. disaggregation). So, you may need to create algorithms that generate the data you need (data synthesis). At NET2GRID, we have high-quality data resources, the first step to success.
Regarding data generation, there have been some great breakthroughs in the last few years in the domains of computer vision and natural language processing that can be used for time series analysis problems such as the GANs and the VAE architectures and, of course, the transformer models which are the architecture behind most of the GPT AI models.
3. Over the past three years, you have been working as an Energy Expert and Machine Learning Engineer at NET2GRID. What are some of the most exciting projects you have worked on during this time? How have you been able to combine your research on energy disaggregation with the practical side of it at NET2GRID?
Working for NET2GRID all these years has allowed me to participate in exciting projects collaborating with an amazing group of people. As an Energy Expert, I was part of the design of many algorithms, including some serious breakthroughs for the company. The ones I would like to highlight would be the following:
A. The POC (Proof of concept) analysis and results of the DER (Distributed Energy Resources) solution for Powerley which was the first step of NET2GRID entering the US market.
B. The creation of the Deep Learning model that detects Electric Vehicle end uses on the edge in real time.
4. There's a growing concern about the potential dangers of advanced AI and machine learning technologies. Do you think there are real risks that these technologies entail and could lead possibly to increased personalization and societal bubbles, especially when it comes to the future generations?
The breakthroughs in this domain, on a daily basis, are many. New models and novel applications based on AI and GPT networks are introduced constantly. Since new products are constantly being developed and offered to the public, there is the possibility that mankind is not fully aware of the potential risks. This is something that happens with every new technological achievement. For example, the wheel was invented first and then the rules on how to drive it were made. This means that society starts building the rules and the relevant framework on how to use AI products, what is ethical and what is not, what is useful and what is not, afterwards. As Uncle Ben (from Spiderman) said: “with great power comes great responsibility” (and after this, he died and became a legend). I believe that we will find the best balance on how and when to use AI for the best interest of humanity and for the common good.
5. Looking ahead, what is your vision for the future of energy, especially with the integration of AI and machine learning? How do you see these technologies transforming the energy sector and the clean energy solutions available out there?
Ideally, I would imagine a world without any energy waste, where every small amount of energy could be used for something meaningful. Of course, this is the holy grail, but, step by step, this goal could be achieved; the job may be easier using AI / Machine learning along with all the knowledge mankind has developed over the years. After all, AI is a human-made invention, it can and should be used for good purposes. Keep in mind that, at this time we are talking, the AI on its own is just a set of tools that cannot do anything if not given the correct direction.
6. Apart from your research on AI, your side gig involves playing a handful of musical instruments and also having your own band! If you could choose any three musicians, dead or alive, to join your band for a jam session, who would they be and why?
Yes, that is correct, music is my passion. I have been studying and playing music since I was ten years old. My main music group is a jazz duet (voice and guitar), but I also play with other musicians too, mostly blues and rock music. Your question is very difficult to answer since I love more than one style of music but if you insist on this, here are my top three deceased musicians with whom I would love to jam. Bill Evans (jazz legend, pianist), Chet Baker (legend of cool jazz, trumpeter and vocalist) and John Bonham (drummer of Led Zeppelin).
7. Being a full-time dad while managing your responsibilities as a researcher, engineer, and musician sounds like a hustle! How do you balance all these roles and what are your tips for others juggling similar commitments?
The answer is that I don’t, haha! It is very difficult to keep a good balance between having a family, a job, and other activities. What keeps me standing and doing my best is the support from my family and wife combined with my passion for creation and solving problems. I would suggest prioritizing your family first and then doing what excites you more, without much compromise.
Comments