Dr. Katia Walsh

Chief Global Strategy Strategy and Artificial Intelligence Officer, Levi Strauss & Co.

Supernova Award Category: 

Future of Work: Human Capital Management

The Organization: 

Levi Strauss & Co. (LS&Co.) is among the world's most admired companies, the original San Francisco start-up, jeans inventor, and pioneer in Environment, Social, and Governance impact. The brand is a lifestyle icon at the center of culture, innovation, and democracy. Levi's has been worn by miners and farmers, Marlon Brando and Albert Einstein. LS&Co.'s “profits through principles” mission stems from its founder who funded the University of California first scholarships, with half of them for women . Since then, LS&Co. has been defending civil rights, supporting the LGBT community, giving a voice to immigrants, fighting gun violence, protecting the environment, and promoting voter registration.

The Problem: 

LS & Co. aspires to be the world's top apparel company, famous for its values and brands. In 2019, it started to build an-industry-first global Digital, Data, and AI capability that integrates digital products, big data, machine learning and automation. The vision is to deepen our consumer experience, enhance internal efficiencies, and transform the company from analog, manual, and intuitive to digital, automated, and precise. We've been able to attract some sophisticated talent thanks to a powerful brand, global footprint, and rich image data for computer vision applications. However, newly hired data scientists and machine learning engineers face a steep curve in learning the intricacies of the apparel industry. That entails everything from design to product development, sourcing, manufacturing, merchandising, inventory management and planning, distribution, logistics, Wholesale and Direct-to-Consumer sales, eCommerce and physical stores. To be succesful, external hires also need to integrate in the company through a keen commercial focus, clear communication, and relationship management. This takes at least several months, even for the most talented hires. At the same time, a digital transformation cannot depend solely on external hires. It can only happen if it penetrates the entirety of the enterprise, involving people who know the company, have credibiltiy, and become agents of change.


The Solution: 

As part of our digital transformation ambition, we created an inaugural Machine Learning Bootcamp, an immersive training in coding, machine learning and agile ways of working, uniquely designed for LS&Co. employees, using LS&Co. data to solve LS&Co. problems. This was an 8-week full-time hands-on virtual program teaching Python, libraries, exploratory data analysis, time series, neural networks, computer vision, natural language processing, recommendation systems and deploying models in production. We took employees out of their jobs to train them. Though coding programs are not new, this is a pioneering effort for employees with no prior experience at all in coding and statistics. Instead, we looked for the aptitude and attitude to master data science: Problem-solving, curiosity, and the perseverance required in the scIentific method. This democratizes the traditionally rarified AI field, while training employees across the enterprise.

The Results: 

We received 360 applications from all countries and functions around the world. To ensure objectivity and fairness, while honoring the democratic nature of the program, we designed a highly-selective but NO-CODE application process, including essays, two challenges, Hackerrank, and interviews. We selected 43 finalists across four virtual classrooms in North America, Europe, and Asia, representing 14 different locations outside of the company headquarters. The democratic nature of the application process paid off: 63% of finalists were female, with a median tenure of 2.5 years representing 14 locations across corporate offices, retail stores, distribution and data centers. Despite the challenging curriculum, every participant completed the program and is now a certified machine learning practitioner. A quarter of the graduates joined the Strategy & AI function, while the other three quarters returned to their functions, where they apply their new skills to automate, digitize, and optimize our processes and operations throughout the enterprise. We will hold another bootcamp in the fall and, by the end of this financial year, train more than 100 employees in machine learning. The democratic nature of this program attracted the attention of The Wall Street Journal, which featured the program as the only one in apparel retail: 



While it is too early to have all specifics, the graduates are already contributing to the value delivered through Digital, Data, and AI at LS&Co.: $76m per year, with a 76% gross margin, and an EBIT of $33m. LS&Co.'s CEO Chip Bergh said at graduation: "This is a big step for us in our journey to digitally transform the company... What I'm excited about is we've now taken a cohort of people who have experience in the company and in the apparel industry. And we've now built those skills with all of you, and this is going to supercharge our data science organization."

In the graduates' words:

Ji Yong Yi, senior planner in Merchandise Planning - "One of the biggest values I believe I can take back to my team is being able to speak both langiages - data and AI, and Planning - and be empowered to accelerate our digital transformation at LS&Co." Dani Jefferson, a distribution center employee - “After this bootcamp, I’m excited to help figure out ways to solve some of our bigger business obstacles and create inventions that will help everyone else.” Leen Allaert, product development director: "[This has] also been a personal transformation where you get the hunger for learning more and making a change. I hope I can bring the technology and appetite for making changes into the team to apply technology where we can to make life easier." 

The Technology: 

Course Outline

Week 1: Exploratory Data Analysis with Pandas, NumPy, Matplotlib, and Seaborn

Week 2: Data Science Process; Decision Tree Models, Ensembles

Week 3: Feature Engineering, Agile Methodology

Week 4: Time Series

Week 5: Neural Networks

Week 6: Natural Language Processing, Sentiment Analysis, Word2Vec, FastText

Week 7: Recommender Systems, Collaborative Filtering, Autoencoders, GANs

Week 8: From Notebooks to Scripts, Data Apps, Dashboards

Disruptive Factor: 

This is revolutionary for a number of reasons: 1) teaching sophisticated skills to people with no prior experience in coding and statistics - especially in a creative industry, not seen as analytical or technologically-advanced; 2) taking valuable employees out of the business for 8 weeks - a big disruption and not immediately supported by managers who did not want to lose their people; 3) enhancing diversity in a traditionally male-dominated field; 4) deploying graduates back to their jobs - instead of hiring them into a dedicated centralized organization; and 5) delivering concrete benefits as part of the instruction and afterwards. While this is not easy to manage, it will, in the long run, ensure a better adoption of Digital, Data, and AI throughout the enterprise. Each of the graduates is now transforming their job, and with that, the entire company: A store manager is using her skills to predict the inventory her store needs to stock; a distribution center operator is predicting which equipment is likely to fail in the next few weeks; a factory quality assurance employee is identifying the attributes that predict a product will pass quality inspection; a designer is inputting classic art, e.g., Van Gogh's "Starry Night" to create new clothing designs; a product developer is optimizing the costs of production, and an HR team is automating diversity, equity, and inclusion reporting

Shining Moment: 

Message from a bootcamp graduate who returned to her job in Planning: "I wanted to thank you! I’ve been at Levi’s for almost 9 years and it is by FAR the best thing I’ve done. I’ve been interested in data science for the last several years, and was thinking about taking a sabbatical from my job at Levi’s to take a data science course. I loved everything I learned in the bootcamp, and have been very interested in bringing what I learned into my day job."

About Levi Strauss & Co.

Levi Strauss & Co. is one of the world's largest brand-name apparel companies and a global leader in jeanswear. The company designs and markets jeans, casual wear and related accessories for men, women and children under the Levi's®, Dockers®, Signature by Levi Strauss & Co.™, and Denizen® brands. Its products are sold in more than 110 countries worldwide