Machine Learning for Time Series Data With Dirk Michiels
Machine learning is the evident proof of today’s futuristic society. Time will come and it will be a common term just like artificial intelligence. Joining Chad Burmeister is the CEO of Tangent Works, Dirk Michiels, to discuss how they utilize machine learning for time series data. He explains how rapid forecasting helps different industries come up with accurate and effective strategies. Dirk also talks about the role of machine learning in creating explainable results, making the work of data scientists way easier.
Listen to the podcast here:
Machine Learning for Time Series Data With Dirk Michiels
I've got Dirk Michiels with me.
Thanks for having me.
It's great to have you. Dirk is the CEO of Tangent Works, a machine learning company. We'll dig into a little bit more on the types of companies that they help. We're talking AI so let's roll up our sleeves, dig in and understand a little bit more about Dirk's company and what he's working on. Dirk, welcome to the show.
Machine learning can be used for full-costing, classification, and anomaly detection.
Thanks, Chad. We in Tangent Works are focusing on machine learning on time series data. Time series data is a pretty difficult animal to crack. We see it in all kinds of processes around the world whether it's energy consumption, the use of products supply chain questions and also in the agricultural culture sector. All those things are time series related. If you want to create machine learning work before this, it’s pretty hard. Data scientists know that. We have hyper-automated this process completely and made it into an augmented machine learning process that can be used for full- costing, classification and anomaly detection. I hate the term machine learning because, for me, it goes about the use case that you're acquired. What is the business value that is agreeing to give you? That's where we differentiate.
We tend to focus through our network performance on the value that delivers to our end customers. We've made the solution extremely simple to use. Our solution is available to a REST API. It can run a workload on premise and a lot of devices as well. The neat thing is that we also make it available from various platforms that you already have in your IT environment ranging from data platforms such as Microsoft Azure Data Factory, Alteryx, also things like Amazon Glue. Those also would fit into the visualization with Qlik, Power BI or Tableau and that makes it not only easy from a data standpoint to use but also easy to integrate with your IT landscape.
How did you get into this business? You're in Belgium. Is that where you were born and raised? Where are you from originally?
Machine Learning: AI and ML will not be treated as separate things in the future. They will become part of day-to-day life.
We were born and raised in Japan. One of our co-founders was a professor at a university in Japan focusing on Information Geometry and Information Criteria which are two branches of Mathematics that we happen to use to do the feature engineering that was required for trans for full-costing. This technology allows us to do a forecast extremely fast and accurately but also creates an explainable result. That's where the company comes. Later on, we moved back to Slovakia where we have an office. We also have one in Brussels, UK, near New York and in Dubai.
When you were growing up in Japan, what were the types of things that interested you? If you were to say, “My passion was blank.” “What were you passionate about when you were younger?
I joined as a CEO later, I was one of the initial investors in the company but the thing is that Jan Dolinsky who is our CTO was extremely passionate about Information Geometry and Information Criteria and saw the link to machine learning. As I said before, the company went on the journey from being a pure machine learning company to a company that was focusing on delivering business value for end customers. That's where the machine learning field is moving to. In a sense, we are celebrating in 2021 the 50th year of Box-Jenkins who invented the first time series forecasting models. In those days, we could already talk about machine learning but we call it Box-Jenkins. The passion that this company has is to deliver business value from the machine learning data that is available.
Machine learning technology allows users to do a forecast extremely fast and even create an explainable result.
One of my earlier roles in 2000 was with i2 Technologies that ended up being acquired and they were dipping their toe in the water on machine learning and forecasting but they were a little ahead of their time. I'm not sure where that technology ended up. Another good friend of mine has a double PhD in Supply Chain and Operations. He ended up doing this kind of work for a lot of very big organizations so I'd love to put you in touch with Dr. Das sometime. He's a great guy. Does the sales team leverage artificial intelligence in anything? You're using it inside of your products and services to your customers. What about from a sales side? Does anybody in sales leverage any AI in their outreach?
The question is what is AI? I know that's also what we’ve been talking about, that you are doing an amazing job in this lead generation thing. Our solution is very much focused on time series forecasting so the time series is not sales-related. We can make use of this going forward. AI and ML will be an integral part of our business anyway. We will not treat it as a separate thing that is innovative and that people must use. It will become of our day-to-day life. In a few years' time, we will be talking about AI, ML anymore. That's also the reason why we invest hard in setting up our partner ecosystem because we feel that if we can work with people that understand the business of our customers and that's where we can make a difference because they understand the end customer's needs, they see where predictive analytics can come in and this can be greatly supported by machine learning.
I'm looking up where Dr. Das is. He now is at Saatchi & Saatchi as the Chief Data Officer. He's been there since 2020. He would be of interesting contact. Before that, he was on the Advisory Board at Harvard Business Review, Viacom, Amazon and McKinsey. He's a thought leader in the AI space. This is an exciting conversation, Dirk. I thank you for coming on the show. I’m very much looking forward to continuing the dialogue. If you're not familiar with the organization, check it out Tangent.works. They're looking for partners. If you're in the sales arena and work with companies that have needs like CPG, transportation and a number of other industries, feel free to reach out to Dirk. How could they get ahold of you if they're interested in becoming a partner, Dirk?
Machine Learning: Thanks to machine learning, it is easier to predict the amount of energy a country can consume or the time needed to fill the tank of an E-vehicle.
You can go to our website and you can look under “Partners,” there's a tab, “Find a Partner.” Apply there, we'll get in touch and talk about our value proposition, give you a demo and also see what the joint business propositions are. For end customers, it’s the same thing. Our website is pretty intuitive so you check out all the information there and reach out. There's even a free trial of our solution available so it's easy to connect with us.
I would think these opportunities can be quite large even into the seven-figure size.
There's a large diversity of applications. We are doing things in estimating the number of vehicles that are required in the leasing company, for example, up to predicting the loaves of bread that the bakery needs to prepare. We are predicting the amount of energy that is going to be consumed in a country, the price of commodities, we are doing solutions in eVehicles where we are predicting how long it will take to fill the tank of the eVehicle and optimize this in the energy that is available. Lots of use cases are available.
It makes me think of Pete Buttigieg who's one of the people in government now. They proposed to start taxing cars on miles driven rather than taxing the gas in the gas tank because of the move to electric vehicles. There's wear and tear on the road so I'm sure there are some interesting political arguments on both sides of that. It’s getting the data and understanding who would be impacted by that and what could be the result in the additional income for the government around taxes if they move to that. Those are some of the things that you could help with on the big data and analytics side.
All of those things are also driven by changes in legislation. This is different around the world. This is where our partner ecosystem comes in big time because those people are often very much connected to what's happening in each individual country.
Dirk, thanks for joining the show. I appreciate it. I hope you enjoyed the conversation. We'll catch you on the next show. For now, signing out. Cheers.
Thanks for inviting me, Chad.
Thank you, Dirk.
Important Links:
Dr. Das Dasgupta – LinkedIn
About Dirk Michiels
Experienced General Manager, Leader with a solid track record in building Information technology (IT), System Integrator and Independent Software Vendor organization and driving revenue and profit in BENELUX and global context.
An executive for IT solutions in Utility Sector, Public Sector, Industry, a broad experience in system integration (project and service management), solution provider, product development (independent software vendor on Microsoft Dynamics and Oracle Platform) covering strategy, business development, delivery.
Driving leadership for strategic changes and transformations - developing a company from 8 to 400 people.
Worked in BENELUX context and global space with focus on India, APAC and US.
Experience with developing a worldwide partner network (Microsoft and implementation partners) for a leading Independent Software Vendor (ISV) solution on Microsoft Dynamics for the utility sector.
Motivated by the creativity to develop the strategy and coaching people to drive execution.
