Are you building infrastructure or developer tools to power AI-first technology products? Head over to Air Street Capital.
The Research and Applied AI Summit (RAAIS) is a community for entrepreneurs and researchers who accelerate the science and applications of AI technology. In the lead up to our 5th annual event on June 28th 2019 in London, we’re running a series of speaker profiles to shed more light on what you can expect to learn on the day!
At RAAIS we have a focus on translating cutting edge technology and research into production-grade products for real-world problems. To that end, we’re excited to welcome Travis Addair from Uber Technologies! Travis is a Senior Software Engineer working on the Michaelangelo machine learning platform (read more here). Michaelangelo covers the end-to-end ML workflow to enable internal teams to seamlessly build, deploy, and operate machine learning solutions at Uber’s scale. The system supports traditional ML models, time series forecasting, and deep learning.
In particular, Travis is a contributor to Horovod and sits on its Technical Steering Committee within the Linux Foundation. Horovod is Uber’s open source distributed deep learning framework for TensorFlow, PyTorch, Keras, and MXNet. A sneak peek into the system's performance: in the figure below, you can see that training with Horovod lets a user take full advantage of their GPU resources to achieve 90% scaling efficiency across hundreds of machines at once. Using Horovod, users can unlock the full potential of distributed training with just a few lines of Python code added to their existing training scripts.
In his talk, Travis will share the story of how the Horovod project got started and how it solves the distributed training problem. He will provide comparisons with alternatives, give us a demo of the product and share real-world case studies both at Uber and outside of Uber. Travis will also share recent performance and accessibility improvements, as well as the roadmap for the project.
Prior to joining Uber, Travis has worked on a range of exciting applied ML projects:
Automated maps generation using deep learning @ Google
Developing data-intensive distributed systems used to analyze several hundred terabytes of seismic data to detect underground nuclear weapons testing @ Lawrence Livermore National Laboratory
Map creation from images of extra-terrestrial worlds @ U.S. Geological Survey
Mobile games development @ Storm8 and Super Evil Mega Corp
Travis holds an MS in Computer Science with an emphasis on AI from Stanford University and a BS in Computer Science and Mathematics from Northern Arizona University.
Welcome #RAAIS2019, Travis!