Risk Data Scientist Spotlight: Joan Kirunga
Meet Joan Kirunga, a risk data scientist at Stanbic Bank Uganda. She develops and deploys machine learning models to anticipate inherent business risks using customer conduct and transaction behaviour data. These models include a credit risk model to predict customer loan delinquency, financial loss forecasting for the business using historical data, and ATM cash demand forecasting to inform ATM replenishment, among others.
Tell us about your path into data science. How are you using AI and data science in your work?
My path into the data science world commenced with my Master’s degree in Information Technology at Carnegie Mellon University, where I majored in data science. During my pursuit of my Master’s degree, I was offered a data science Graduate internship opportunity with Liquid Telecom, where I was able to put the data science lessons I learnt in the classroom into action. I replaced their manual reporting by designing a Microsoft Azure data warehouse to store and ingest semi-structured data from sales, customer, and finance databases and building a dashboard visualising their key performance indicators.
Upon completion of my Master’s degree in 2019, I continued on the data science path, taking on a role as a senior data analyst at Bboxx, a solar energy firm. Here, I conducted pilot and research studies, analyzed the performance of solar products, and forecasted product failure, showing the probability of product failure at different timestamps while in use by a customer.
What inspired you to pursue a career in this field? Is this what you’ve always wanted to do?
Growing up, I was a huge fan of a geeky cartoon called Dextor, whose intelligence and love for science fascinated me. I always thought I’d be an electric or telecom engineer, but in high school, when I was exposed to a computer during our computer studies classes, I was captivated and wanted to explore how they worked. This desire led me to pursue a Bachelor’s in Software Engineering at the prestigious Makerere University. While my academic background provided a solid technical foundation, my interest in the realms of AI and Machine Learning seemed like an unattainable dream only for those overseas. Nevertheless, driven by my curiosity, I embarked on a self-guided journey, starting with the powerful language of Python. As I delved deeper into the world of programming, my enthusiasm for data science and AI grew, and I dedicated myself to continuous learning in these captivating fields.
What are some of the key trends or developments in AI and data science that you are most excited about, and how do you see these shaping the future of the field?
There are several exciting trends in AI and data science that have the potential to shape the future in different ways:
I. Reinforcement Learning will continue to make an impact in the field of robotics and automation by enabling machines to learn from their environment or processes and make
decisions based on continuous training or trial and error. This trend has the potential to
completely change our working dynamics and reduce human error in our everyday tasks.
II. Ethical AI has become paramount with the increase of AI integration into various domains such as manufacturing, healthcare, and banking. There is still a strong need for efforts
to address bias and promote accountability in AI systems to ensure privacy, security, and fairness.
III. AI in the health care and financial sectors will change our livelihoods. In the health sector, lately there has been a lot of information about electronic health records, which will aid in AI-driven diagnosis, medical imaging, and disease prediction models. On the other hand, in the financial sector, there is an ongoing trend of AI fraud detection through natural language processing and image recognition, which is fundamental with the increase in digitalization.
In your opinion, what are the key challenges facing the AI and data science sectors? What needs to be done to make the sector more inclusive and impactful?
I. Data privacy and security are paramount in the AI and data science industries. AI integration into healthcare, finance, manufacturing, and the justice system is loaded with the fear of misuse or mishandling of personal data. Therefore, there is a need for data protection and governance measures to mitigate risks. Considerations: The growing use of AI in critical decision-making processes, such as healthcare, criminal justice, and finance, has raised ethical concerns. Issues related to privacy, transparency, accountability, and the potential for AI to exacerbate societal inequalities need to be addressed to build trust in AI technologies.
II. As a data scientist, our biggest challenge is data scarcity and quality. In many domains, acquiring high-quality labeled data can be expensive and time-consuming. Data scarcity can limit the performance of AI models. Prioritising and investing in research can somewhat mitigate this challenge.