Conversations #75: Sijing Zhu, Senior Manager Data Science at The Home Depot

Women Who Code Conversations 75     |     Spotify – iTunes – Google – YouTube – Text

Akshita Korwar, Software Developer at Hulu and Network Director of Women Who Code Seattle, interviews Sijing Zhu, Senior Data Science Manager at The Home Depot. They discuss the importance of continued learning, listening skills being a priority for success, and time management being key in accomplishing goals.

Tell me a bit more about your journey and how you became a senior data science manager.

I am currently pursuing my master’s in computer science with a machine learning specialty from Georgia Tech. I have a master’s in industrial engineering from Georgia Tech and a bachelor’s in industrial engineering from the University of Missouri. When you look at my education and my position, they look different, but they are all data related. My original background was in industrial engineering, where I learned all the statistics, process improvement, and basic machine learning models. My computer science degree focuses more on advanced machine learning models and how to develop databases and sound systems. My whole career path is all about data as well. Before I joined Home Depot, I analyzed data with Siemens and TravelClick. When I joined Home Depot, I started as an analyst and supported our home services marketing department. I got a promotion as an analytics manager. I use all kinds of tools and data, for example, data visualization, statistics, time series, and machine learning to support our online marketing, online, and CRM teams. Recently I was promoted to senior data science manager. This role is focused more on machine learning and data engineering.

What motivated you to get another master’s in computer science and specialize in machine learning? 

When I started my job at Home Depot as an analyst, I was one of the few on the team with some data background. At that time, we didn’t have a centralized data warehouse, and we didn’t have automated reports. We had so many isolated systems for different business lines. Without good data, it’s challenging for me to use all my skills. We started with building a SQL database to centralize all the data sources and build all the automated reports on the top of the database. It’s pretty fragile, and it fills almost every week. Our business is still amazed by the work I’ve done. I was demoing a report to our VP. We combined all the data sources from different business lines and systems. It’s interactive and automated. Our VP then used his pencil and calculator to cross-check the other numbers. He was so surprised that our numbers matched and made sense. My report brought new insight into the business they had never had before.

I got to skip a level with a promotion to be an analytics manager. My biggest project then was leading a group of enterprise warehouse engineers to move all the home services data to the cloud. We also designed a way to connect all the upstream systems. For example, all the web analytics systems to our downstream system, our audit system, and our lead management system. The project involved many teams. Data engineering teams and developers were supporting all kinds of apps and systems. There were also business partners involved. That’s when I found that my knowledge is not enough. There is a significant knowledge gap between engineering, design, and computer science. At that time, no one on the team knew all three areas. I wanted to fill the void. I applied for a computer science degree at Georgia Tech. I took classes like database design and software development. Those classes helped me understand how to design a good table schema, why we need to avoid more values, the software development life cycle, how to design and test the system, and why we need version control.

I also took advanced modeling classes, deep machine learning, and AI. I have learned so much, but sometimes it’s still unclear how to make everything work in a real-life situation. I met my current director. He has the reverse learning path as mine. He has a computer science background but learned data science later. He built all kinds of foundational data systems and ML flows. All of those systems are pretty reliable and scalable. I said, “Wow, that’s exactly what I want to learn. I hope I can work for him.” I moved to his team two months ago. From my experience, it’s hard without a blended knowledge of engineering, data science, and computer science. You won’t be able to design a comprehensive data solution and integrate it with all the business systems to solve business needs. I feel like I decided to pursue another computer science degree.

How are you balancing both being a data science manager and a student? Does Home Depot have any special programs supporting their employees while they are at their job and pursuing a degree? 

Time management is the key. I always like to plan everything. I know my day job is already busy enough. I look at my schedule and see that I probably can’t spend 10-20 hours weekly on school. I can probably take one course every semester, which means I can graduate in around 3-4 years. I also give myself some buffer. I will skip a semester if I’m busy in a few months. I try to finish all my work during the day, and the night of the work days, I spend around 1-2 hours on school. On the weekend, I spend one more day on school stuff. So with this plan, I kind of still have plenty of time. I can still go to bed around 11:00 PM every day.

I have a free day during the weekend too. I watch TV, read books, and play with family and friends in my spare time. So it’s just an average person’s life.

Home Depot gives me massive support for my schoolwork. Our company encourages all its associates to continue learning and pursue new degrees. We have a tuition reimbursement program. Georgia Tech has pretty high-quality and affordable online Master’s programs in both computer science and analytics, so I don’t have any financial concerns. Also, most of the data science team at Home Depot is remote. That gives me extra hours per week on all the commute time. Our leaders and my team are super supportive as well. There are a lot of my coworkers that are currently taking another degree. Our leaders value this learning and believe it can enhance our skills and benefit the company.

People often talk about data science, machine learning, and artificial intelligence almost interchangeably. Could you tell us how they’re related and what’s different about each subset? 

Machine learning is building an algorithm that can learn independently using all the historical data. AI is broader than machine learning. Artificial intelligence is more about making a computer system that can mimic human intelligence. The key difference is that AI does not require all the historical data. Also, it can include additional components, for example, robotics. Data science is a broader field. It uses data and tools and techniques to find patterns, derive information, and solve real business problems. For example, using statistics, visualizations, models and algorithms. It overlaps with AI and ML, but they have their area.

What is one technical innovation in AI or ML that excites you the most? 

We have developed a really good data foundation for home services. I’m excited about our next step: building a scalable and reliable machine-learning system. Currently, we are building a lead scoring model. It’s not just one model. It will support hundreds of different categories for home services. It will serve all kinds of different business needs. For example, how can we prioritize our outbound phone calls? Which customers should we re-engage with first? There are so many use cases and we are trying to design it in a way that it’s auto-scalable. Once the model figures out what category and the business case, it can automatically find the related data feature and automatically select the model and tuning. It will find the best model itself. For this, it requires machine learning, data foundation tables, a seamless data pipeline, a scalable machine learning computing engine and also algorithms to automate all the model selection parts. Eventually, we will integrate it with other business systems to combine all the components to create a reliable and seamless flow.

What is an important quality as a manager that you think is required to make the team thrive? 

I learned all my managerial skills from The Home Depot. They provide many in-person and online training programs to teach me how to be a good manager. I also learned a lot from our internal leaders who I interact with every day. From my previous experience, I feel like one of the essential qualities of being a good manager is listening. You need to listen to your associates’ needs. What are their career goals? What type of work do they enjoy the most? What are they good at? This will help me assign them the right path and help them grow in an area they’re passionate about while also achieving the business goal. This keeps my team motivated and growing fast. Also, listening to business partners is essential. Different teams have different perspectives on the same problem. When you’re launching a new project, operation teams probably care about performance and training. IT will care about security concerns and resources. The data team will care more about data accuracy, model performance and SLA. Finance will care about self-impact and cost. All of them are important to consider to make the project successful. It also helps to balance the potential conflicts. It’s also important to listen to your leaders. They normally know more information, have more experience, and see the bigger picture.

What is something that you enjoy doing outside of your work? What do you do to unwind? 

I love cooking Chinese food. It makes me feel relaxed and peaceful. My cooking skill is pretty good. When the dish is ready, it smells so good and tastes good, making me happy. Also, it’s a hobby that my family and friends can enjoy too. I love seeing their happy faces when they like the food. Sometimes I apply my process improvement skills on cooking. I try to plan all the steps ahead before I make the dish. I try to minimize the cooking time and the number of cookware I have to wash later while not sacrificing the taste of the food. Besides cooking, I also like traveling. I list crazy things I want to do, like skydiving, exploring caves and riding a dolphin. I use my statistics mindset to decide whether to do it or not. For example, I will search the death rate and compare it with the car accident death rate to decide whether to do it.

What’s one pro tip for Women Who Code? 

Set up your goal and never give up on learning. I know there are a lot of challenges in being a woman in tech. I’m not even talking about bias here. I’m fortunate that I never had any sex bias in my entire career. I’m talking more about women’s natural roles and for example, giving birth to a child and taking care of your kids. I have a three-year-old daughter, a day job in tech, and night school. I can still balance this through time management. Don’t feel afraid of any challenges. If you are willing to do it, you can do it.

***************
Guest: Akshita Korwar, Software Developer at Hulu and Network Director of Women Who Code Seattle
Producer: Kimberly Jacobs, Senior Communications Manager, Women Who Code