Insights In AI and Data Science (Episode 7): Q and A with Vangelis Oden, Data and Intelligence Engineer at Indicina Technologies (Podcast Included)

Insights In AI and Data Science (Episode 7): Q and A with Vangelis Oden, Data and Intelligence Engineer at Indicina Technologies (Podcast Included)

Single Speaker - Vangelis.png

🕔Read Time: ~11 minutes

Hello, community!

We understand these are trying times the world-over but we hope you and yours are staying safe from the virus. Remember to follow recommendations from recognized health authorities in your region and help stop the spread of the virus.

If you are on a budget, we also feature an article on how to make your own mask here using only a piece of cloth, needle, and thread.

We are back to our "Insights In AI and Data Science Q 'n' A With Our Experts" blog-post series. You can check the previous episode in this series on following good methodologies with Emmanuel Okorie (Data Scientist at Maon Technologies) by clicking here.

These "Q 'n' A" episodes will help you get to know various practitioners and experts in the field of Artificial Intelligence and Data Science at a personal and career level; these are mostly non-technical, insightful, and fun conversations.

Listen To This Interview

If you prefer listening to this article, you can click the section below to navigate to our podcast page (make sure to subscribe 😉).

Check the table of contents below to quickly jump to a section in the article.

Table of Contents

  1. Who is The Interviewee For This Episode?

  2. What You Will Learn From This Interview?

  3. Main Points We Got From The Article.

  4. Personal Questions (Get to Know Vangelis).

  5. Applied Machine Learning/Data Science and Career-Related Questions.

  6. Vangelis's Advice to You.

You can learn from Vangelis's workshop on "applied classification in machine learning and how his organization implements classification solutions" with the Port Harcourt School of AI. You should follow our Twitter page and subscribe to our YouTube channel to watch other insightful expert-led sessions both nationally and globally.

Who is The Interviewee for This Episode?

In this episode, we interviewed Vangelis Micheal Oden who is a Data and Machine Intelligence Engineer at Indicina technology, a Fintech startup that provides smart services for their clients. Vangelis shared insights about his life, education, career, qualities of a competent Machine Learning Engineer and Data scientist, challenges he has faced and how he overcame them. We were able to pick his mind and brain so to say, lol. This interview has lots of useful tips and insight that can go a long way for you as you build your career in Data Science and Machine Learning.

What You Will Learn From This Interview?

  1. How Vangelis started a career and the challenges he faced, including how he overcame them.

  2. How to develop your Data Science job-search and interview-preparation strategies.

  3. How small businesses can effectively come up with good data and AI strategy to work with, and why they shouldn't consider AI technology first.

  4. Qualities you should have to be a very good Data Scientist or Machine Learning Engineer.

  5. Why you should get started learning Data Science and AI now.

  6. How Machine Learning can improve sectors in African countries and what to take advantage of in the process.

Main Points We Got From The Article

Tip from me to you.gif

Via Giphy.

Here are the main points we got from this episode with Vangelis;

  • Small businesses or organizations now have a variety of accessible tools that can help them capture and leverage the right data and perform profitable analytics with these tools.

  • As a start-up, using logic systems first-off can help you get started on how you can leverage these AI and Machine Learning technologies for your organization.

  • Proper understanding of an organization's challenges and a focus on the job descriptions required can help bolster your chances of landing an AI or Data Science job.

  • Learn methodologies and principles of data science rather than the tools so that you can be dynamic and flexible in applying these methodologies through tools and technologies. Know the technique, think scientifically, and this will aid your creativity and innovation

  • Being a team player is crucial for your growth as a Data Scientist or ML Engineer.

Interview Section With Vangelis Oden

Personal Questions (Get to Know Vangelis)

Stephen: I’m so glad to have you here today, welcome Vangelis. Please can you introduce yourself?

Vangelis: My name is Vangelis Micheal Oden. I hold a Bsc. in Electrical Electronics Engineering from the University of Port Harcourt, Choba. I hold two master's degrees in Data Science and Machine Intelligence. I’m currently based in Rwanda but I work for a fintech startup in Nigeria, Indicina Technologies. I work there as a Data and Machine Learning engineer where I basically design cool stuff for the company.

I would say I’m really down to earth and really cool lol. I’m all for designing solutions, building things from scratch, and using models and new technologies to affect people’s lives, that’s why I went into engineering. I’m crazy about new technology. So that’s me in a nutshell.

Stephen: That’s an awesome introduction, I can testify to you being cool haha. What did you want to be when you went to college? Did you go to college with a mindset of solving problems with engineering or did you want to do one thing and then end up doing another?

Vangelis: So before getting into the university, I always wanted to break down systems and solve them. So I was more interested in computer science and hacking. I wanted to exploit systems and I needed a course that could help me wrap my head around this. The obvious choice would have been computer science but I realized that I couldn’t learn it to the capacity I wanted here in Nigeria, so I decided that the next best choice was Electrical Electronics Engineering because it was close enough to robotics and my future goal of helping computers think.

Through my undergraduate course, I found some of my best courses to be control engineering, I just enjoyed the ability to give logic to a system and watch it perform just that logic. That helped me gain the intuition on teaching computers to learn, not just teaching them to act but also teaching them to learn to act.

So that was my feel through learning Electrical Engineering. After my undergraduate, based on the environment I found myself, Port Harcourt being a city of oil and gas. The next line of action would have been to employ my skills working in the oil and gas sector on the systems they already have, but I wanted to take it a step further. I wanted a place where I could build systems from scratch to finish so that was why I moved away from Port Harcourt and the oil and gas.

Stephen: That’s awesome. Your journey has been exceptional and intentional right from the start. That speaks a lot of volumes. I’m curious though, you mentioned how you were already interested in AI, but were there key moments that spores your interest in this field? Of course, you didn’t know about Data Science and deep learning and the rest from the start.

Vangelis: Okay, first of all, I’m a very big fan of Sci-Fi movies. Though when you rapidly get into Deep Learning you realize that most of those movies are just mumbo jumbo. Most of the movies I was watching at the time opened my eyes to the possibilities that lay in untapped data. Movies like i-robot turned me in that direction. Otherwise, it was basically me speaking with mentors to understand where technology was headed at the time. Having just graduated from university, I needed to know my place in the scheme of things, was going to be an inventor or a builder or a researcher. After speaking with some mentors on what could be the frontier in technology I decided it was worth it giving data a chance based on how untapped the resources were.

Stephen: That’s a really awesome inflection point. Pretty much 80% of our guests have talked about how Sci-Fi influenced their passion for transforming what could into what should be by turning them into Engineering problems that can be solved. Great insight! Thanks for answering the questions in this section, Emmanuel. We are going to move to the next section where we have a conversation on your career and career-related activities. Stay tuned, readers.

Stephen: Hi, readers! Welcome to this section of questions where we explore Vangelis’ thoughts on Machine Learning (ML), Data Science, and his career journey. So pretend I’m not a tech person, can you explain Machine Learning and Data Science in simple terms for me?

Vangelis: I’ll explain this and then I’ll give an example. According to a friend of mine, this is a rhetorical question if I say “All politicians are corrupt, Obama is a politician therefore he is?” The obvious answer in your mind is a generalization you have been able to make due to experiences overtime. Now, Machine Learning/ Data Science is the science of understanding facts that have happened over time and using them to generalize on a new concept or theory in the future. I think this is the most layman way I can explain Data Science.

It takes historical data and draws inference to predict a future scenario in any type of problem classification to the prediction of time-variant, continuous variant, and discrete variant quantities.

Stephen: That’s perhaps one of the most unique takes on this particular question. Based on your own words how do you think Data Science and Machine Learning as individual fields aid problem-solving?

Vangelis: The use of Data aids problem-solving in several ways. People often think when they’re getting into Data Science that they’re going to just be predicting things. I have seen in a lot of Small and even large organizations that most of the data Science being done just relies on the ability to understand what the data is saying concerning any particular problem and to make decisions based on this.

For instance a loan company, yeah one of the interesting things that can be done is to build credit scoring models that can predict default for a customer, another use of data could be just to understand how users interact with your application forms from this you can rearrange your forms so people can apply for loans in the shortest possible time. Here you don’t even need a prediction model. You can build a better workflow based on data.

Data Science encompasses all of the analytics from looking through data and understanding patterns, Machine Learning for prediction, and Deep Learning to deal with different variants of data. All these different fields form the main area of data science, and for anyone out there you can get into any of these areas and excel in them.

Stephen: That’s a wonderful take on the broad areas that Data Science can be useful both on a small and large scale. You’ve what Data Science and Machine Learning is to the audience. Now I’d like to know what’s one thing about Data Science/Machine Learning that is clear to you but is not so to other people (engineers and non- engineers alike)?

Vangelis: A lot of people feel that when they get into Data Science more of the work is going to be around building world-class models and really robust Deep Learning systems but most of the work actually lies in Data preprocessing in reality. The truth is that you’ll spend 80-90% of your time preprocessing Data. When you do a lot of MOOCs (Massive Online Open Courses) they give you already cleaned data to do minute cleaning and get done quickly. But in the real world, sourcing your own data, understanding of the data is useful, preprocessing, and cleaning the data are some of the issues you might face. There are no standards to preprocessing so you have to formulate how to approach each particular dataset you’ll be having access to and that’s the most interesting part of Data Science.

Once you can wrangle data in different dimensions then building a model is an easy task that just requires reading a couple of papers and documentation on the problem you’re trying to and before you know it you’ll have a working solution.

Stephen: Yes, a lot of people are always under this delusion of just having to build world-class models. But if you think of it there are a lot of tools that automate model building but very few that automate preprocessing and I think individuals and organizations need to take note of this. Having worked with small and large companies, how would you advise a small company to adopt an AI strategy? What would they need to be successful at AI and data science?

Vangelis: A lot of companies feel that they need to jump into AI and Data Science because it’s available. After speaking with a lot of small business owners and startups that are thinking of adopting Data Science, one of my advice to them is to build kick-ass logic systems. By logic systems, I mean massive IF statements, that will get things started. After some time you can grow into that AI capability. If your problem cannot be built as a logic system then you cannot successfully say that it can be modeled as a Machine Learning problem.

I always tell business owners to think about the future, consider what you want your Data Science team to look like, and start building the capabilities before you get there so you don’t have to eventually start having issues with overhauling the system. This is one of the issues that we have with many of the commercial banks in Nigeria, they did not optimize for the science of data. So now finding themselves in a fast-driven AI world they can’t adapt quickly because their systems were built for old business methodologies. Data is as good as useless if it is not being analyzed in the right way at the right time and presented to the right individuals in authority.

Stephen: I really like the approach of going from traditional logic systems to a more Machine Learning driven approach to solving problems. As you have rightfully claimed, if it can be solved by logic then it can possibly be solved by AI. One thing I’d like to add is conducting that needs analysis before deciding if a problem can be solved with AI or not.

How would you advise small African companies to develop a proper data collection strategy to prepare for a more data-driven AI-powered future?

Vangelis: I love Google's model of tools for data collection, data warehousing because they are more optimized for startup companies to easily analyze their systems. I was speaking with the CEO of FarmstoYou, a small startup that runs an application that helps farmers in Sub-Saharan Africa and all she does is plug in her application to the google analytics system and it helps her get information on how and who is interfacing with her application. It’s an easy thing to do, you can google how to move data through Google Cloud Platform from an application, it’s easy to set up.

However, if you want to build a more primary system then that’s left to your pocket. On google cloud, you get charged based on how many times you query your data and not based on how much data you store. It helps not to worry about the volume of my data but to concentrate on the quality and how I can make sense of the data, this is something that should be of importance to startups.

Stephen: I really like the idea of using Google Cloud tools even though google is not the sponsor of this interview and neither are we endorsing them. I personally have been using Google Cloud Platform for more of my jobs, the big query ML services allow me to build fast insights at the basic level and also store large amounts of data at unprecedented speed as well as querying and getting analytics on data, this helps support small businesses that are looking for ML solutions. Startups taking advantage of Cloud Platforms is one thing I would recommend because cloud vendors also take care of data security. The next question is more about the challenges you faced when you were starting out in this field.

Did you face any of these challenges? If yes, how did you overcome them?

Vangelis: For me becoming a data scientist in Nigeria was really really difficult because at the initial time nobody was doing ML in Nigeria, a couple of people were doing analytics but nobody was doing the deep areas I wanted to go to. I wanted to classify images, make inferences from texts and speech, and being in Nigeria made it a little difficult. Luckily, Data Science Nigeria came and I attended a boot camp there in 2017. I was shocked at how much there was to know there, I had a couple of courses on Data analytics, moving and analyzing data but I did not know where to go from there and knowing what to do next or learn next is always key to how you are able to grow.

After what was happening in DSN, I knew what my target was and I just kept on firing in that direction. I knew I wanted to do Machine Learning, to understand how to analyze and recognize patterns, I did not know what algorithms to use for what problem at the time to get the best results or how to deploy and maintain those solutions. I realized those were the things I needed to add to my arsenal of knowledge.

So after DSN I kept firing in different directions, I joined my first Data Science job in 2017 and from there it just went on. In the field, once you’re able to tackle real-world problems, build models and maintain them and see how they affect business and how it translates to profits and risk management then that’s it. Understanding the big picture was what I wanted and so I stayed with a Financing Investment for some time and joined KPMG before leaving for Inidicina that had been my growth over the years.

Stephen: Oh that’s a really awesome experience. And to laud the amazing efforts of Data Science Nigeria like I would say the highest impact technology non-profit in Africa and how they have managed to shape a lot of careers in Data Science is not just the nation but also globally on a massive scale. Talking about skills and qualities, what qualities do you think are most important for someone to become a very good Machine Learning engineer?

Vangelis: One of the most important is your analytic skill. Before you jump on a problem you have to be able to understand where the data is coming from and where it’s going to before you start building the models. You have to think like a scientist, be extremely analytical and creative in thinking around these things. Once you can understand and are creative enough to come up with a solution the next thing is your programming and debugging skills. They need to be top-notch because you going to experience bugs, bugs from hell, bugs from everywhere in different areas of your system from the programming to the deployment to the refactoring and if you do not understand the language you’re using and how to identify your problems, then you’ll always find yourself on maybe Stack Overflow or Google.

Many of the times the main work happens in your head even though to someone watching it may seem you spend all your time in front of your laptop. To me these are the three main skills, of course, there are other skills like soft skills, leadership, team building, negotiation skills.

Stephen: One skill that you purportedly mentioned in the scientific skill, having to craft out those hypotheses and also prove them on a scientific scale and it’s one thing that I see a lot of ML Engineers ignore because they are more about let’s hack away and it will work. All models work but not all of them are useful. Scientific thinking is one thing I think people should learn and it’s about asking a lot of questions and making sure your assumptions are verified.

Vangelis: Data scientists are questionnaires. You have to keep asking questions until the solution goes out. The truth is you are not a product leader, you are a product follower so you need to ask all the possible questions as to how the particular product you are building improves things in the long run. So you have to always be inquisitive of systems, data, and processes.

Stephen: Thank you so much for that take. An important question; how do you think ML can best help Nigeria improve and how can the country leverage the technology?

Vangelis: I feel the only way we can leverage this is by accepting this as a viable solution to some of our problems. Before Software Engineering became a rave in Africa, we wouldn’t readily accept this because not many people would trust it, and it was also the problem of accessibility and reliability.

A lot of people are still skeptical of AI today, even the government. Once governments are able to understand that these systems are not bad and can help predict a lot of situations that will happen in the future and this can help us take proactive measures to mitigate those scenarios or build more robust systems to utilize positive results of the prediction. Only then can we see a lot more of AI flourishing in Africa.

I always take the example of how China has moved their society to an AI dependent society such that you can just walk into a shop and pick anything you want to buy without money and you get automatically debited based on your face, your social score and the kind of data the government is tracking and the solutions being provided. I think the government needs to will have to start these applications where they see them, it should be looked down on and over scrutinized because it’s AI.

My second take on this is that AI can be done by anybody, it’s not restricted to geeks. There are different spectrums and areas in this field. You can plug into any part of your interest. I would implore anybody who is interested in how to build systems that can understand data to just go ahead and start building you’ll learn the maps as you go. The more people we have in the field, the more an ability to expand the field of AI, algorithms to use, and methods. That’s why I was very happy about the master's Machine Intelligence I took that was designed for Africa by Africans.

All hands should be on deck, everyone should be interested in gathering data. There are a lot of things that should be done and I implore everyone and anyone to just get in and get your hands busy with something, fill those gaps and you’ll be able to make it anywhere you choose.

Stephen: That’s a fascinating take, and to draw from your first points one of my favorite quotes; “... enough of history can tell us about the future”, that’s one thing I know for sure anything that will happen in the future has a historical transient. When it comes to the government leveraging this technology we should assume we’re in the parallel universe (😁) not trying to be insulting.

On this technology putting people out of business, what is your thought on how it affects work in the future especially for Africans who might not be ready for this revolution?

Vangelis: My major thought about technology is that it augments how we live and interact with our surroundings. It gives us more leverage and time to think about other things. AI and ML are not new concepts, they’ve been around for some time the only reason they were not being used the way they are now for years was because of the availability of data and computational resources.

Admittedly in every augmentation cycle of technology, a lot of people lose jobs, this is not due to any particular bias. Think about it from the perspective of the business owner who would rather reduce operational expenses in terms of salaries and at the same time increase revenue, every business person wants to stay on the profit side. I have the option of employing more people in the field of that machine and paying them the same salary I was giving to my workers and getting a better profit. So I’m simply changing my workforce and not necessarily taking away jobs. Take the pandemic happening for instance, it has made people with the skill to work remotely more relevant because of the circumstances.

So I would say yes AI will come and take away jobs but it will also create more jobs that center around how it is being managed, operated, and improved. So you owe it to yourself to either go home and get yourself suited for the next workforce or go home sit down and lick your wounds while mourning how obsolete you’ve become. I implore people to increase their skills and not look at the constraints.

Even the AI systems get obsolete in a few months and have to be retrained so that means an AI engineer is always needed to improve the systems, so if the AI nOt is going to be Learning then you also have to learn.

Stephen: I think I would commend what technology has done to technologists and Engineers, it has made us have what Carol Dweck called the growth mindset. It has become part of us to upskill or self. So I would encourage anyone out there to have this growth mindset that’s the only way to survive this new era.

That leads us to one thing we’re excited about, what are you excited about in the future of Machine Learning and AI?

Vangelis: One of the areas I’m excited about is the evolutionary learning of ML systems, it is Machine Learning systems taking the best qualities of themselves and self-improvement in the next iteration of the model. For instance, I learn on a particular image review recognition task, the best way I am able to learn as a model is to adapt that and send my best weights to another model that can use that weight together with its own best weights in another task. This way systems are able to understand themselves and evolve over time into new learning systems that understand a wide variety of data types and problems. So evolutionary ML is one thing I’m excited about.

The next is generative adversarial neural networks. Being able to design new systems and identities from noise data and being able to make them look so real that even AI systems can’t detect the difference between these created systems and real-world systems, things like created images, videos, platforms that even advanced systems cannot differentiate. I’m also fascinated about the things I don’t know in AI yet, a lot of things I know now, I did not know them 5 years ago so I’m excited that there could be things in AI I don’t know about and I could even possibly be the one who invents or discovers this. This keeps me open and ready to solve all types of problems and hopefully one day creating a solution.

Stephen: That’s awesome, maybe you can give us some links to research on evolutionary Machine Learning for the viewers.

Vangelis: Sure, I’ll get it across.

Stephen: What would be your advice for ML Engineers looking for a data science job in the country?

Vangelis: It's easy to say I’m looking for a Data scientist role or an ML engineer role, but you need to understand the dependencies of the roles and skills you need to excel in these roles. For anyone starting out in Data Science I’d say take some time to understand how to fetch data, look at data and understand how to identify patterns in data, then take a data analyst role and spend some time it depends on the industry you find yourself, try to understand what exactly you’ll be doing as a data analyst and spend some time in a place where you are really doing the work. It always helps you to understand the feel of data and why the data are acting the way they are acting.

Once you start interacting with data it will be easier to relate with the models and the results that will come out from the models. This makes it easier to interpret your results to management for decision making. So I would say take a Data Analyst job and do it for some time and then from there migrate gently to other fields like ML, Deep Learning, and other areas. You can’t become a Data scientist without understanding data.

Stephen: What would be your advice on how to prepare for an interview?

Vangelis: Most hiring managers are looking for people who can be flexible. Take up roles in data analysis, and then business analysis. Another thing we look out for is a researcher and creative thinker, ready to challenge methodologies to help to bring in better ideas.

Also, someone who doesn’t just understand the tool but also the methodology behind it. Scientific thinking to understand different types of problems, so that you can easily adapt to different types of tools and easily derive insights from them. Lastly, team play, you need to be able to work with people in different areas of the field or other fields to churn out multiple solutions in the business.

If you do not know how to work with these people based on your own field you will always end up feeling more or less superior to them and thinking that the work depends on you forgetting that they all play a role in providing solutions. You end up being joined by the team and company and live basically in your space.

Earlier in my career, I made the mistake of thinking I could do it all but then I realized that even a simple suggestion from someone, not Data Science could help my time and resources and get the job done well. You need team-building skills, the ability to work with various types of people.

Stephen: Awesome, talking about thinking in methodology and not in technology is something I think people should take note of, that’s why it’s always important to always best to learn the foundation and principles to things.

Vangelis's Advice to You

Stephen: How do you manage to keep track of the latest happenings in AI?

Vangelis: I think that’s one of my shortcomings. I’m very bad at keeping up with events but I’ve been trying to work on that using an application called meetup. Meetup is an application that you basically see community meetups in and out of your area that you can pay to see.

I try not to get warped into many of those meetings though because different speakers come from different industries with different methodologies, of course sometimes it helps keep your scope open to some of these methodologies but other times it can really trail you off the work you’re doing where you need to implement some particular methodologies repeatedly because they are currently the best methodologies that work currently pending future growth. So that was my head thinking not attending too many of the meetups but if you’re able to segment all that information and you also have the funds to travel around then I encourage you to go ahead and do that. The good thing is that it allows you to network with a lot of other scientists improving and increasing your chances of moving across different companies and different systems including doing some ad hoc projects for people while not working directly with them. I would say I have benefited from going for meetups so it all boils down to you and your needs and career growth in data science.

Stephen: That’s a good thing. Your final words on this, what is one golden advice you would give to beginners looking to become world-class Machine Learning engineers or Problem solvers?


In one sentence, learn the Math. It’s easy to build solutions to get the information online but to improve the system you need to know the math behind how it is operated. When AI Engineering learns math.

Stephen: Thank you so much for your time, Vangelis. Once again our guest on today’s interview is Vangelis Oden, Data and Machine Learning Intelligence Engineer at Indicina. We’re so excited you came to share knowledge with us.

Well, that wraps it all up, readers! Thank you for reading this wonderful interview with Vangelis. We hope you learned a lot from it.

You can take action steps now by giving back to the community: Remember to share your thoughts with the community through the comment section below or on Twitter by sharing this article and tagging us @PHCSchoolOfAI along with points you learned.

We will hopefully keep releasing sessions like these before each of the (#pmlcourse) classes with a guest speaker/facilitator. If you would like to ask Vangelis a question directly, you can reach out to him on Twitter @thatData_guy.

If you also enjoyed this Q and A with Vangelis, do leave a reaction 🔥 to the story, hit the like 👍🏽 button, and share it 📩 with your friends that may be interested. See you soon!

You can also subscribe to our podcast show to get the first-hand notifications on new episodes we released. You can also leave a 5-star rating on iTunes if you enjoyed the show so others can be able to find it.