Engineer to Engineer: Innovative Leadership in Machine Learning
Written by Sapphire Duffy
Sapphire Duffy is an Associate AI Engineer at Kainos and the Director of Women Who Code Belfast. In this edition of Engineer to Engineer, Sapphire meets with Huiming Qu, Senior Director of Search & Recommendation & Data Science at The Home Depot, to discuss adapting to changing technology, staying productive, and the importance of diversity and inclusivity in the workforce.
Can you tell me about working as a technical leader at Home Depot?
I lead a product management and data science team at The Home Depot that supports onsite search, recommendations, and personalization for the company’s digital channels and other data-related aspects of the business.
What was your introduction to machine learning?
My experience with machine learning goes back to my last year of earning my Ph.D. I majored in data management with an emphasis on real-time data processing, such as data transmission in-stock exchanges, balancing the trade-off between speed and quality.
However, my last Ph.D. course was in data mining with Professor Christos Faloutsos from Carnegie Mellon University. I fell in love with it immediately. I enjoy mining immense data sets and applications to find patterns. My first job in IBM research furthered my interests in data mining in various applications, including social network analysis, recommender systems, finance optimization, and supercomputer resource management.
What would you say is the influence of emerging technologies like AI, machine learning, and service architecture on the way engineers approach building products and services for users?
I believe that the core benefit of computer science is automation with speed. We are now capable of reducing friction in the shopping experience and automating various tasks that customers once had to perform manually. Our technology reduces the number of steps in the buying process, saving customers time.
The convenience for our customers manifests in several ways. For example, we rank the most relevant products in higher positions on listing pages so customers can discover products easily. Additionally, we recommend appropriate products or services to help customers complete their projects. We also help customers design rooms with product styles and color matching. Our mission is to help all of our customers 24/7 proactively and seamlessly.
That’s fascinating. What does a typical day look like for you, and what skills and qualities do you need to do your job well?
Time during quarantine looks very different from how it used to for almost everyone. People are tackling more home improvement projects than ever before, so it’s my team’s job to make their shopping experience as seamless as possible. It’s fulfilling to help add a little more happiness, peace, and convenience to people’s lives during this difficult time.
In terms of skills, I think the most valuable is aligning everyone with a common goal, regardless of the situation. This skill is not unique to tech, but it’s an essential leadership element that requires motivating people with distinct talents to march toward the same place and at the same pace. Nothing gets done if everyone is moving in different directions.
(Huiming speaking at the 2019 Women Who Code DataPy Summit in Atlanta, GA)
On that note, do you have any productivity tips?
I’m a huge fan of productivity culture, so I look for those all the time! Most recently, I was exploring how to build better habits that eventually become second nature. For example, I put little hearts on my calendar as a small way to reward myself whenever I achieve a goal, like going to bed at a decent hour.
Besides sleeping earlier, my other goals — especially during quarantine — include finishing my exercise before 8 o’clock and going for a walk with my kids every day. I’ve been pretty successful at it thus far; I don’t even need to think about including the new pieces of my morning routine.
The pandemic disrupted everyone’s routines, but the thing about disruption is that it provides an opportunity to revise whatever wasn’t working before. I would advise others to be as specific as possible about what habits they form to save the mental energy of making decisions about what to do every day.
I love it. Switching subjects, what do you think is important for companies to do to have a diverse workforce and an inclusive environment?
That’s a great question. Diversity and inclusion are critical for any company because they are ultimately key to finding the right talent. Big and small companies alike want to have people with unique perspectives working for them.
Many members of my leadership team, for instance, come from other parts of the world. The ability to attract a broad range of talent, regardless of where they’re from or where they went to school, amplifies when you have the right people in hiring positions.
Inclusivity also promotes a positive work environment. Every organization should be cognizant of where they source talent from. Using the same channels repeatedly yields the same kind of people who don’t benefit the company as a whole and leads to stale workplace culture.
This awareness helps my team at The Home Depot to tap into many diverse talents. In particular, I think women offer a specific and unique perspective of leadership — but if you don’t have women at any level, you can’t grow these potential leaders to the next level. I like to hire women and then groom them for leadership to keep this positive cycle going.
What has been your biggest challenge, and how did you overcome it?
My biggest challenge is one I’m still confronting: I’m a perfectionist, at work and in life. You should see me plan a trip; I’ll plan out every detail of every day with itineraries and decision trees. I’m also prone to over-preparing for presentations and not resting until everything reaches the “perfect” stage.
This habit isn’t helpful because one, it’s exhausting, and two, it diminishes my return on investment — or rather, my return on labor. The first 80% of a project is essential, but if you spend countless hours on the last 20% that you could have used for something else, then the overall return is much less than it should be.
So, I still have room to grow, and I have to remind myself not to do certain things.
Can you recommend some resources for people looking to become a leader in the machine learning or AI spaces?
There are so many resources available that it’s hard to pick. Making a habit of continual learning in AI and ML is essential because they’re growing so rapidly. It’s also vital to align your education with where you want your career to be.
Everyone learns differently, whether it’s reading books, listening to podcasts, taking training courses, or perusing websites. I prefer listening to podcasts and audiobooks because I like major educational meals instead of snacks. I’ll set aside three days to take on something that I really want to finish or a particular deep learning forum. The critical thing is to make learning a habit.
Do you have any tips for the WWCode community?
Women and people of other gender identities always have lots of different things pressuring us, but we also live in a time that confronts us with unprecedented challenges. My advice is to take this opportunity to stretch yourself and acquire new skills. Learn to enjoy the moments when you have to do something you don’t want to and turn the experience into something positive.
___________________________________________________________
Huiming Qu is the Senior Director of the Online Data Science & Platform team enabling Search, Recommendation, Real-time Personalization, Visual shopping, and various other innovations for The Home Depot’s digital channels.
With more than ten years of experience managing large AI and Data Science programs, she has a proven record of driving billion-dollar contributions with scalable machine learning solutions and strategic innovation.
She has a Ph.D. in computer science from University of Pittsburgh, holds six issued patents with more in filing, and more than fifteen academic papers around data management, machine learning, and optimization.
