Tarry has over 20 years of experience working with data and has advised CxOs, Leading Government Authorities (including Ministry and President Level) of global organizations and country states to setup data-driven organizations from scratch.
He is a regular visitor in Finland and has advised & trained some of the largest corporations here. Check also his upcoming class on AI & ML!
Now we had a unique opportunity to interview Tarry.
How did you end up for a career involving machine learning and deep learning?
I was always interested in AI since the beginning of the 90s. While I have always been in the involved in data driven projects from the beginning of the 2000s the real breakthrough came in 2012 when I was involved in a project to develop a social analytics platform for a startup. We used to call it data analytics platform, who knew that soon that would be called NLP (natural language processing, a form of AI which is getting extremely popular today. Fast forward to today, I am glad that I continued to keep my skills updated in AI.. I have personally been involved with some really exciting projects – both large and midsize, that I could have never done if I did not dive into Artificial Intelligence a few years ago.
What have been the highlights of your career so far?
I have been fortunate to work both on the data as well in the management side of my career.
The key highlights are as follows: I have led data driven projects for enterprise clients, small and medium (SMB) enterprises and also with startups. These customers have been in various industries such as retail, banking, recruitment services, oil and gas, telecom, healthcare and various other industry verticals. The most important thing that all these projects involved was that in the beginning they were not mature to have a full visibility of what their data could mean for them but it was great to see these departments of companies transform gradually into data-driven companies.
In our trainings for managers, I constantly share the best practices with managers and engineers with the help of AI driven use cases.
How many people have you trained?
So far, since 2017 we have trained over 27.500 people worldwide in AI. These are -IT professionals, managers and executives and even CxOs.
What is the best thing about training people?
While there are many platforms online that provide machine learning and deep learning trainings, personal and immersive trainings have very specific purpose and goals – especially in Covid times where people increasingly feel alone and don’t know what their real skills are and where are they in their journey to become an AI expert. I primarily enjoy the interaction that I have with the candidates and we share our knowledge and experience which normally is not available online.
Secondly, I am personally involved even after the training is over. I stay in touch with the candidates that we have trained so even after the training we are in constant touch with each other and keep discussing the problems, issues or potential job opportunities that arise for our trainees to do something more with the data that they have.
Since the launch of LiveAI (https://liveai.eu) we have aggresively expanded our learning by doing methodology evne further. Our learners have gone to create their own companies, they have secured jobs with leading AI startups and some have even joined our DK labs.
This is what I always wanted to do: make my network successful!
Many people think there is a hype around AI, what is your opinion?
I recently wrote an article about fake AI versus Real AI and it highlights the factor about how much AI been hyped by people who do not understand it. According to one research there are at least 40% of companies that say that they do AI but they essentially do not have anything AI inside their systems platforms or processes. Media reporting too can be very hyped up and generally makes people wonder what is real and what is not.
Fortunately, the reality is even more pleasing. The startup and investment ecosystem is growing dramatically and within a few years those who would have invested in AI could have a dramatic lead and those who would have ignored this would really fund it hard to survive in a data-driven AI economy.
What is the reality behind the hype?
The hype is primarily due to the fact that it is a popular buzzword, just like what cloud computing was 10 years ago. Every company was a cloud computing company back then! Today you are in the AI hype circle – especially with the NLP mania which is making AI do even more interesting things.
My advice is to continue to focus on your learning path and keep learning about machine learning and deep learning because there is going to be a big change inside your IT environments, in your data structures and inside your overall company. If you prepare now, you will be ready for all these new technologies when the AI economy would have arrived.
For instance, April 28, 29th we are launching our new training called MLOPs – running succesful projects in production.
This training connects your recently acquired AI skills to developing apps and products which you would like to release into production environments. This is what seperates professionals who are doing real-world AI from the hype machine. The training details you can find with our partners Ari Hovi website.
Why a CEO of a big company should be interested in AI?
CEOs of large companies have a lot to gain if they use artificial intelligence inside their business operations but they also have a lot to lose if they do not get data driven.
What I mean by that is that CEO is generally a focus on capital structures, P&L and are generally are leaving their day-to-day operations of managing their data assets to CIOs and CDO’s.
This is normally not a problem if you have regular day-to-day operations to run but, in an AI, or data intensive economy, it is very important to collect data to aggregate data into derive value out of data on a daily basis and that is why CEOs of companies need to focus on data driven initiatives and need to manage these data different initiatives also rather actively.
Think as an example: Many companies are running AI projects, but still 85% of these never make it into production because o poor or no MLOPS strategy to deploy and maintain these! Imagine: If a company could save €10M a month using MLOPs deployment, not doing that means it is losing €10M monthly! This is of course not what company’s want.
What is your advice how he/she should lead adaptation of AI within the company?
The five key steps that the CEOs need to take is the following:
- to align and drive their strategy and vision that is based on data nothing else but data.
- to prepare the staff to become real machine learning experts and not hire people who have some random profiles of AI. This only creates an illusion that they are doing something with data.
- they have to continue to focus on developing algorithms and minimum viable product within a matter of weeks so that they can see the results quickly
- Plan ahead to build a defensible business but use by unique multi model an algorithm strategy – what I mean by that is companies need to realize that it is building a competitive advantage with multiple algorithms and your own data driven models is going to keep you way ahead from competition
- Finally, strengthen your lead by combining hardware, software and services models.
Do also please read my article from last week that goes deeper into how to achieve these steps.
How the CEO could make sure the employees have the understanding and skills to proceed with AI?
This is a really good question! Many CEOs believe that their employees have the understanding and skills to proceed with data just because they have done some data project inside their company but this is a wrong assumption because many employees who may be developers, database administrator’s or some other technical competencies are not prepared to actually do any projects in machine learning or deep learning because it requires totally new competencies and skills.
To make sure that these employees can transition and start becoming machine learning or deep learning experts, CEOs need to invest their time and efforts to make these employees familiarized with artificial intelligence and its general concepts.
The CEO also needs to make sure that the managers are equally aware of the opportunities that artificial intelligence can bring to the company. By merely preparing engineers and technical staff to become experts will not help the company.
It is the right combination of the business managers and the engineers, that will make AI successful inside the company. If CEOs wish to succeed in the AI economy and develop new products and solutions based on AI, then they need to invest in both managers as well as engineers.
What kind of training programs you have provided for companies?
We provide several kinds of training programs for companies:
First, there is an executive training program which helps the CEOs, CFOs, CEOs, CMO’s and other executives to understand the benefits of AI. This is very essential and executives get visibility into the competitive ecosystem around them, focus on making the right choices and finally aligning their strategy and vision to become a data driven AI company.
Second, we deliver training program for their managers — these could be sales managers, marketing managers, R&D managers or business managers. We help the managers understand the benefits of AI in the form of intensive workshops that focuses on identifying solutions inside their own business domains. This helps these managers to understand about the date of opportunities that exist inside the domains and also helps them have a meaningful conversation with the engineers.
Third, we train the engineers with all forms of machine learning and deep learning techniques such as computer vision, natural language processing and all other forms of latest machine learning techniques. By learning these techniques, the engineers are prepared to answer questions that are driven by the management by developing algorithms, models and potentially solutions that can be integrated inside their existing services.
Finally, we also help these companies develop prototypes, algorithms, models and even software solutions based on AI. This is the final and the most important step in advising companies to become aware of the opportunities for machine learning.
What are your advices to an individual who wants to start an AI career?
Many beginners either take on too much and get overwhelmed or are just not able to start because they see this too intimidating.
My advice to individuals who want to start a career in AI would be to pick up basic skills in AI, do a small project in AI, then pick up intermediate skills and do a medium size project and keep skilling up in this manner.
Your next training is on 28th and 29th this month. Why should you attend and what will you learn?
Together with Ari Hovi we already deliver training in Machine Learning with Python. Here you will learn the fundamentals of the latest AI tools and techniques. In this training you also learn about real-world projects that we have done so you get a picture of what all is possible once you learn about it. Our focus is constantly to provide you a healthy mix of theory and practice so you always go home with a great learning experience which you can apply to your industry.
But starting 28th we are excited to launch our MLOPs training which will go into detail about taking your project idea from incubation to releasing it in a PAAS (platform as a service) or IAAS (Infrastructure as a service) platform. For more details please look at the course curriculum on Ari hovi site.
What will be the next coolest solutions that the companies and individuals will see with progress of AI?
The next wave of great solutions and products have actually already stated entering the market. AI startups are already focusing on robotics and machine learning techniques to help manufacturing companies improve their warehouse operations. Banking and Financial services have already made some progress with fintech but they too will expand as more and more deep learning (NLP) will be used to develop sophisticated solutions. Today language models (NLP) are suddenly getting very popular and especially with GPT-3 (https://medium.com/letavc/apps-and-startups-powered-by-gpt-3-976c55dbc737) a lot of young startups have started developing solutions for the market. Same opportunity is there for enterprises as well to further improve their solutions and services with these AI tools.
Healthcare too will be a growing market: today we already see a lot of diagnostics tools based on AI, soon we expect end-to-end solutions that will ensure that basic shortages of healthcare professionals and tools/techniques will be solved with AI algorithms and models.
Also Energy is a huge market where we will see some ground breaking discoveries in developing hybrid and state-of-the-art renewable technologies. They will all be powered by machine learning – from research discoveries through deployment in production environments.
Where do you see yourself in 2025?
I know this is a heavily loaded question since I think we are just at the beginning of AI transformation. 2017 I was one guy with a dream and today I have our AI labs established in Netherlands, Toronoto, Canada and by end of this year also in Austin, Texas, USA.
We have indeed achieved some really awesome models in climate change we call CAELI and we are soon to launch a ClimateAI startup out of deepkapha.ai labs. For healthcare, we’ve already secured, won and are close to delivering our solutions to the European Union.
And Covid-19 made us rethink our survival strategy – and that has led to the birth of LiveAI platform which are developing continuously.
So I think I am more busy with what am I going to do today and tomorrow and that is the best way to define the future 😊 ( I believe this is what Peter Drucker, the famous management guru said as well)
Thanks Tarry for the interview!
Tarryltä on tulossa 28.4 myös uusi MLOPS Training – Running successful AI projects in Production -koulutus.
Aihe on maailmalla todella suosittu ja useat yritykset haluavat kouluttaa henkilökuntaansa aiheesta. Tarry sai pelkästään viime viikolla aihetta käsittelevästä webinaarista useita 10-100 hengen tarjouspyyntöjä.