How can you identify a zebra if you’ve never seen one before?
Data scientists know the answer: Learn how to identify horses and know that a zebra looks like a horse with stripes — the classic example of implementing zero-shot learning within an artificial intelligence model.
Data science is one of the top three jobs in tech and one of the top ten jobs overall, according to U.S. News and World Report. The scorecard rating speaks for itself: Data scientists report high salaries, a flourishing job market, good work-life balance and high opportunity for growth. But with myriad opportunities, data scientists have the opportunity to work on vastly different projects in an array of industries.
So in a field full of options, what makes a job stand out as a zebra among the horses?
According to data scientists at Nordstrom, Inspire and DRINKS, cutting edge tech, meaningful projects, collaborative colleagues and room to grow make the difference. Read on to learn what they shared with Built In LA about what they love about their projects, work and team culture.
Built In LA learned one thing data scientists want to know about Nordstrom’s legacy retail brand: their tech stack. “Nordstrom is powered by languages like .NET, C#, JavaScript, PHP, Python, Scala and Swift; React for libraries; and Node.js, Play and Symfony for frameworks,” Built In LA reported last year.
What is the coolest project you’ve worked on as a data scientist at Nordstrom?
One thing I’m working on right now is building an AI-driven product attribution platform. The work uses machine learning to assign features to our products, which can in turn be used in various downstream applications such as search filters, assortment planning and styling. We use representation learning to create a vector space for our products based on images, raw text and various other categorical data. Using zero- or few-shot learning, we can then assign attributes to products quickly and at scale. This platform gives us the tools to put the best and most relevant products in front of our customers as well as the agility to adapt to evolving customer needs.
What excites you about the work you’re doing?
As we continue to grow our product assortment at Nordstrom, inspiring digital product discovery and driving personalization at scale becomes critical. The attribution platform gives us the flexibility to assign product features at scale while still allowing us to continually refine our taxonomy to give our customers the best possible shopping experience. By using the state-of-the-art computer vision and natural language models, we can get a comprehensive view of our products with basic product metadata and, in turn, give our customers the same Nordstrom in-store experience but within a digital context.
One thing I love about working with AI in fashion is that it is constantly evolving.”
When do you love most about the culture of your team?
One thing I love about working with AI in fashion is that it is constantly evolving. Trends, fashion events and styles are in constant flux — making it a rich area for state-of-the-art research coming out of universities and consortiums. Being able to connect those emerging ideas to our products and ultimately to our customers is an essential part of meeting our customers on their terms, and I love that it gives me a chance to constantly evaluate new technology and build creative solutions.
Inspire has a lofty goal — pushing consumers toward a clean energy, net-zero carbon future. According to Greg Chu, director of machine learning engineering, the ambitious work has also benefited him individually. “Besides being a critically valuable project for our company, it challenges me to grow on a technical level as well,” he said.
What is the coolest project you’ve worked on as a data scientist at Inspire?
The coolest project I’ve worked on is electricity forecasting. Knowing the amount of electricity our customers are going to use, both in the short and long term, is critical to our company and the customer experience: the forecast tells our trading desk how much energy to buy from the market, it informs our operations teams about how valuable customer cohorts are via lifetime value calculations, and it is central to how we price customers on our customized, monthly subscription product for unlimited, 100 percent renewable energy. We get to experiment with some of the latest in time series modeling as well as large-scale data processing technologies.
We get to use cutting-edge technology to solve impactful and meaningful problems.”
More generally, what excites you about the work you’re doing?
We get to use cutting-edge technology to solve impactful and meaningful problems.
We are a team of full-stack data scientists that get to develop using modern data tools like Airflow, dbt, Snowflake, Streamlit, AWS, Kubernetes, Spark, Databricks and more. We aim to use the most accurate and up-to-date modeling algorithms while also prioritizing simplicity and practicality — all with the aim of scaling our impact on the environment and on our customer experience. To date, we’ve helped our customers collectively avoid over 5.6 billion pounds of CO2e from entering the atmosphere.
When do you love most about the culture of your team?
The data platform team here is one of the most inspiring, fun and kind groups of people I’ve had the honor of working with. My favorite part of our team is the continual commitment toward personal growth and learning and the culture of mentorship that accompanies that. Developing talent is a leadership value that Inspire aims for across the company, and I think the data platform team does this well.
There are so many ways this shows up in our culture. We have team members present external topics of interest to the group in the form of a reading series, attend conferences together and present deep dives on our work. Maybe most significantly, there is a willingness to work through problems together, whether that be brainstorming system designs or pair programming development.
DRINKS is reimagining how businesses connect with consumers in the alcohol marketplace. Their platform relies on artificial intelligence and machine learning, which drives their data science team to innovate, according to Blake Hunter, senior director of data science.
What is the coolest project you’ve worked on as a data scientist at DRINKS?
One of the coolest projects our data science team is working on is what we are calling “Perception AI.” We are building models and insights to uncover how internal and external wine properties, national trends and historical behavior drive customers’ perceived value of wine. Leveraging tools from machine learning, mathematics, natural language processing and computer vision, our team has built a suite of data products focusing on a range of perception areas from visual label branding to textual merchandising to historical and rising trends of internal properties of wine.
What excites you about the work you’re doing?
Because of the complexity and difficulty of the problem, very few companies are trying to uncover the science of selling. It is exciting to be part of a company pushing the edge of innovation. Like how Amazon transformed DTC and Netflix ushered in streaming video, DRINKS wants to revolutionize how wine is sold online.
DRINKS has built a culture that feels like a family, where everyone has a voice and we can do anything together.”
When do you love most about the culture of your team?
Even though the majority of the last few years have been remote, DRINKS has built a culture that feels like a family, where everyone has a voice and we can do anything together. DRINKS has enabled this inclusive culture through a flexible working environment and organized meetups. Regular outdoor meetups have provided an opportunity to integrate deeper with your team, connect across groups, build new partnerships, celebrate accomplishments and even chat with the CEO.
