Sep 9–12, 2019

Deep Learning coming to the Tire Industry: Forecasting warehouse staffing needs with LSTMs

Tianchu Liang (American Tire Distributors)
4:50pm5:30pm Thursday, September 12, 2019
Location: 231

Who is this presentation for?

Data scientist, data engineers, data architects, director of data science,

Level

Intermediate

Description

Deep Learning has been a sweeping revolution in the current world of AI and machine learning; it helps Teslas see the road properly (CNNs), it helps SpaceX lands rockets automatically (Reinforcement Deep Learning), and it makes machines translate better (RNNs). The list goes on and on. But how does this new, hot, technology help traditional industries? What about for a tire distributor company? At American Tire Distributors (ATD), our data science team is rejuvenating the company with machine learning solutions. In this talk, I will go over a warehouse staffing solution, where I utilized LSTM recurrent neural network model ensembled with fbProphet to generate staffing level forecasts and further optimized with CVXPY for maximum optimality of staffing schedules. The solution is now being used everyday across the entire U.S. in 140 distribution centers to cost-effectively staff more than 2000 people daily and on track to realize ~10% in labor cost savings.

The talk will cover the overall business problem context, initial machine learning prototyping, resolving challenges in data, compute, as well as application automation. In the end, I will share my key learnings in developing this solution, including lessons on technical side of things as well as business learnings.

Prerequisite knowledge

Conceptual understandings on deep learning, experience and expertise with python and data technologies (SQL, databases, Hive). Comfortable with cloud computing concepts.

What you'll learn

With open source technologies, cloud computing platforms, current cutting edge deep learning and AI techniques are READY for any traditional industry companies to improve their business processes and create concrete values.
Photo of Tianchu Liang

Tianchu Liang

American Tire Distributors

I am a physicist/mathematician turned computer scientist, and now a machine learning enthusiast. Through years of working as a data scientist, I develop and deploy machine learning solutions to solve real world business problems, such as using LSTM to forecast staffing needs, using xgboost models to execute real-time online customer behavior classifications. We live in an amazing era, where machine learning algorithms conceived some decades ago can be put into reality with a few lines of python code. As one of the first two data scientists in company history to join American Tire Distributors, I helped grow the data science team to a size of 12 within a year; and we are now developing machine learning solutions to help the company in supply chain, sales, warehousing, as well as eCommerce.

Leave a Comment or Question

Help us make this conference the best it can be for you. Have questions you'd like this speaker to address? Suggestions for issues that deserve extra attention? Feedback that you'd like to share with the speaker and other attendees?

Join the conversation here (requires login)

Comments

Alex (Tianchu) Liang | DATA SCIENTIST
04/23/2019 2:04am PDT

Implementing state of the art deep learning models is hard, put those models into production and scale up is harder, getting the business buy-in and put deep learning to generate real business value is even harder. And this is the main challenge faced by non-tech industries when applying AI. In traditional industries, data is messy, tech infrastructure tend to be outdated, and people’s mindset requires change, how do we apply cutting machine learning and push for business impact? Come to this session to find out.

Contact us

confreg@oreilly.com

For conference registration information and customer service

partners@oreilly.com

For more information on community discounts and trade opportunities with O’Reilly conferences

Become a sponsor

For information on exhibiting or sponsoring a conference

Contact list

View a complete list of O'Reilly AI contacts