Pratamamia Agung Prihatmaja: Machine Learning for Personalized Services
As Telkomsel evolved into a digital telco company, the company leverages the benefits of big data and artificial intelligence. In such circumstances, more and more exciting projects arise. One of the projects is the optimization of machine learning. Nowadays, in Telkomsel, machine learning is not only being used to drive business decisions, but it has come to a level in which it has a direct impact on the customers.
So, how does Telkomsel optimize machine learning in its operations? To find it out, we had a chat with Pratamamia Agung Prihatmaja who is part of Telkomsel’s IT Analytics and Machine Learning Department. From the discussion, we will explore the workings of Telkomsel’s digital talents and their role in designing technology that can bring real impact for customers.
What is your role as a part of the IT Analytics and Machine Learning Department in Telkomsel?
I work as a machine learning engineer and I am responsible for deploying machine learning models daily. I work closely with other team members who have their respective roles, such as data scientist and data engineer. Our jobs usually start with the request to make predictions, forecasts, and data insights related to Telkomsel’s business needs.
From there, data engineers will collect data from various Telkomsel’s data sources, such as BTS (base transceiver station) and internal IT systems. Worth noting that we only gather data that relates to the goals of the machine learning model. Then, the data will be used by data scientists to build a machine learning model. The next task will be my responsibility as the machine learning engineer to deploy the machine learning model.
To ensure the workflow runs effectively and efficiently, we implement a way of working that promotes collaboration by actively communicating with each of our fellow team members, especially when we are looking for solutions to solve problems that often arise in machine learning model development, such as mismatches on the data.
How long will it take to build a machine learning model?
It is safe to say that building machine learning models can take a lot of time, depending on the complexity level of the model. Some engineers may do it faster and some others may take a longer time, depending on the situation.
The reason is, when we are designing a machine learning model, we are very dependent on data, and most of the data is obtained from trends that occur in that time. So, at one time, the machine learning model that we created could be obsolete when there were various behavioral shifts from customers. From there, we need to refine the existing machine learning model, then see if it works well and is relevant over time.
What made you interested in machine learning?
I think it is worth it for you to know that in my early career, I worked as a software engineer. So, why did I enter this field? When I started to learn more about machine learning implementation in the industry, there were a lot of new things that I could explore, and I was like “ok, that’s cool”.
If I can compare those two fields, I feel like software engineering is more predictable than machine learning. Also, machine learning can bring out many interesting insights from the development and deployment of machine learning models.
If we could talk more specifically, what is interesting about working on machine learning at Telkomsel?
Many people think that this technology is one of the backbones of Industry 4.0. And if we talk about this revolution, Telkomsel can have a big role in the progress of Industry 4.0. This is because Telkomsel has transformed into a digital telco company, so the company has many use cases of machine learning that can strengthen its role in accelerating Industry 4.0 in Indonesia.
My job at Telkomsel is also interesting because I have the opportunity to provide added value to customers by utilizing technology. Not only that, but I also had the opportunity to build machine learning models from scratch. That way, I can enrich my experience time after time because I would find out ins and outs of developing a machine learning model, including the constraints that can be challenging.
Another interesting thing that I can find from this job is I have to be able to explain this machine learning model to other departments, especially those which are involved in the company’s business growth and sustainability. Therefore, I am grateful that I am not just talking to computers when I work.
Could you explain the use cases of machine learning development in Telkomsel?
The use cases cover various products and services, from MyTelkomsel app to Dunia Games. The application of machine learning in MyTelkomsel app aims to provide attractive data package offers for customers, while an example of machine learning usage on Dunia Games is the recommendation of articles for readers.
I also had the opportunity to help colleagues at by.U, the first digital telco product in Indonesia from Telkomsel, in strengthening data package personalization for customers and analyzing feedback from customers, including feedback that related to feature development in by.U app. The implementation of machine learning at Telkomsel also covers various internal company campaigns and the development of Telkomsel IoT’s portfolio.
What about the potential for machine learning use cases in Telkomsel in the future?
One of the innovation trends for artificial intelligence as the backbone of machine learning leads to its implementation for decision making, and many companies have done it. Telkomsel, as a digital telco company, will also strengthen data-driven decision-making and put artificial intelligence at the heart of it.
Telkomsel has a huge potential to further develop machine learning because it has various channels that could be optimized with machine learning models. One of its implementations can be addressed to strengthening the recommendation system to provide a more personalized and precise customer journey and offer for each customer. Machine learning also can be used in another wide field, from strengthening the company’s data security system to the optimization of virtual assistants.
How about the impact of the implementation of machine learning for customers?
Machine learning can assist us in terms of doing work and make the job easier. Telkomsel understands that, so we use machine learning to create personalized services that can improve the user experience. This has become a real impact of implementing machine learning for customers in the way it is ensuring that each customer service channel can provide the maximum possible personalization.
What are the skills that are needed to pursue a career in this field?
It depends on the role that you choose because there are quite a few roles in the machine learning field. If you want to be a data engineer, you have to study database, pipeline engineering. If you want to be a data scientist, then you have to be very good at math and statistics. For machine learning itself, you need to understand software engineering, how to build a platform, and some knowledge of data science.
I think it is safe to say that many young people in Indonesia want to pursue a career in the technology field. What message can be shared for the digital talents out there?
Never, ever, afraid to try something new, because technology is not rocket science that not everyone can have. Nowadays, everyone can understand technology, no matter the specialization, because there are many knowledge resources available. We can take a major in technology, or attend online and offline courses, even free courses are available everywhere now.
We all can utilize that, because the technology itself will continue to develop, so there is no saying that people who have been involved in this field for a long time will not be overtaken by new people. People who are already in this technology world still need to learn to keep up with the dynamic developments of technology.