Senior Staff Software Engineer Ian Williams from Cruise, Google, Lyft, and eBay discusses the challenges in building self-driving cars, including integrating sensors, ML models, and engineering systems. Topics cover ML perception tasks, safety in ML systems, perception systems in autonomous vehicles, and real-time model challenges in deploying self-driving cars.
Read more
AI Summary
AI Chapters
Episode notes
auto_awesome
Podcast summary created with Snipd AI
Quick takeaways
Building self-driving cars involves integrating various technologies like sensor data and ML models.
Ensuring the safety of autonomous systems requires evaluating machine learning models with statistical approaches.
Deep dives
Challenges in Building Self-Driving Cars
Building self-driving cars involves integrating various technologies such as sensor data, ML models, and software engineering. Ian Williams from Cruise discusses the engineering challenges and strategies in this space, drawing from his experience at companies like Google and Lyft. Perception teams focus on interpreting sensor data to make sense of the environment for the vehicles.
Machine Learning in Self-Driving Cars
Ian Williams shares his journey into machine learning, starting with a background in mathematics and exploring early applications like sentiment analysis using open source tools. He discusses predictive delivery estimates at eBay and speech recognition at Google, showcasing the evolution of machine learning in his career.
Data Challenges in Self-Driving Cars
The generation of a large amount of data by self-driving vehicles poses challenges in storing, processing, and transporting this data efficiently. Ian explains the costs associated with managing and utilizing huge volumes of sensor data effectively to optimize performance and decision-making.
Intersection of Machine Learning and Safety Criticality
Ensuring the safety and predictability of autonomous systems powered by machine learning is a critical challenge. Ian delves into the complexity of evaluating machine learning models for safety, highlighting the importance of statistical approaches and industry best practices borrowed from sectors like aviation.
Autonomous vehicle engineering is a huge challenge and requires the integration of many different technologies. A self-driving car needs data from multiple sensors, ML models to process that data, engineering to couple software and mechanical systems, and much more.
Ian Williams is a Senior Staff Software Engineer at Cruise, and before that worked at Google, Lyft, and eBay. He joins the show to talk about the diverse engineering challenges and strategies associated with building self-driving cars.
This episode is hosted by Tyson Kunovsky. Tyson is the co-founder and CEO of AutoCloud, an infrastructure as code platform. He is originally from South Africa, and has a background in software engineering and cloud development. When he’s not busy designing new GitOps workflows, he enjoys skiing, riding motorcycles, and reading sci-fi books. Check the show notes for more information on Tyson’s work, and where to find him.