.png)
Yoon H. Hwang is a seasoned data scientist in the marketing and finance industries with expertise in predictive modeling, machine learning, statistical analysis, and data engineering. He has several years of experience building numerous machine learning models and data products using Python and R. He holds M.S.E. in Computer and Information Technology from University of Pennsylvania and B.A. in Economics from the University of Chicago.
Currently, Yoon is the Head of Data Science at Automated Data Inc. (ADI). In his role, he leads the development of advanced data science models and solutions for intelligent entity resolution, enabling solutions for automated matching and linking among diverse datasets.
In his free time, Yoon enjoys training various martial arts, snowboarding, and roasting coffee. He currently works in New York and lives in New Jersey with his artist wife, Sunyoung, and a playful dog, Dali (named after Salvador Dali).
We had the opportunity to sit down with Yoon to learn more about his journey in data science, his role at ADI, and his insights on the industry. Here's what he shared with us:
During my graduate studies in computer science, I was fascinated and developed a strong interest in Artificial Intelligence (AI) and Machine Learning (ML). The concept of machines making decisions and predictions based on data and statistical analysis deeply intrigued me, leading me to pursue a career in this field.
I'm a hybrid professional. While AI and ML have evolved into distinct fields of expertise, I started as a traditional software engineer with expertise in integrating AI/ML components into web, desktop, and mobile applications. This allowed me to develop end-to-end solutions, combining AI/ML models with practical applications.
I have long recognized the importance of entity resolution or data matching. In every role I've held, effective data matching has been essential for advanced product development. With the growing volume of data and numerous data sources, linking disparate data is crucial for generating insights and building intelligent systems. ADI's focus on this challenge aligned with my interests and inspired me to join the team, as I've firsthand experience with this need.
I thrive on building AI/ML solutions. The field is constantly evolving, with new technologies and challenges emerging daily. This dynamic environment keeps the work engaging and allows me to create innovative and disruptive models, solutions, and products.
I’ve faced various challenges. However, early in my career, data was not widely discussed or recognized for its business impact. I had to educate others and advocate for the importance of data management and data-driven decision-making. As data sizes and AI/ML models grow, managing infrastructure costs and development remains a continuous challenge.
At a FinTech company in the small business lending space, I developed a solution that accurately predicted customers' financing needs shortly before they arose. This resulted in over $50 million in additional annual loans with just a $5,000 yearly investment, demonstrating the potential for high ROI when solutions meet specific customer needs.
I start by understanding the user's perspective. Identifying their problems, how they will use the solution, and the benefits they will gain is crucial. This user-centered approach has been instrumental in solving problems and developing effective solutions.
I designed, developed, and maintained over 1,000 ML models simultaneously at a marketing tech company. With 10-15 models per client for over 100 clients, the challenge was to create personalized and micro-targeted models for various e-commerce, media, and retail clients while managing daily operations and maintenance. This project allowed me to blend creative thinking to practical, real-world use cases, which was challenging but rewarding.
I've authored three books on practical AI/ML, one of which is used as a textbook at universities globally, including Harvard.
My third book, "Machine Learning and Generative AI for Marketing," is set for release on September 10, 2024. This book provides actionable guidance for leveraging predictive analytics for customer engagement, personalized product recommendations and customer segmentation. It also serves as a guide for creating compelling marketing content using zero-shot learning and more advanced Generative AI techniques.
However, for most marketers, the necessary data to achieve these targeted results is often disconnected or hard to utilize. This is where platforms like ADI come into play, and can help solve data matching and linking problems in marketing. For example, marketers often need to match user session or cookie data with their first-party and third-party data to enhance personalized and targeted marketing campaigns. With ADI's matching platform, marketers can better match users based on different attributes, empowering them to derive deeper insights into user behaviors and further refine their marketing strategies.
Outside of work, I train in martial arts, primarily Muay Thai, Judo, and MMA. My routine involves working, training, eating, and sleeping seven days a week.
In the coming months, I’m focused on simplifying data and entity matching. Although we’ve made progress, there is still room for improvement. My aim is to streamline the entire process from understanding clients’ data linkage needs to selecting and running matching models, making it as user-friendly as possible.
Contrary to fears that AI will replace human jobs, my work focuses on building AI solutions that assist people rather than replace them. The solutions we are building at ADI, while powered by AI, enable human-in-the-loop workflows allowing for human oversight where it matters most.
Large Language Models (LLMs) and Generative AI have significantly advanced human-like text generation, pushing forward language modeling and its real-world applications. I believe this trend will continue, driving further iterations and improvements in LLMs, which will influence the development of new AI solutions and startups.
