07-04, 12:30–12:50 (Africa/Cairo), Hall 2 (Ground Floor)
Maintaining world scale HD maps requires massive compute jobs. We use PySpark and python to map the world, making it performant and cost efficient - as you must at such scale. A glimpse on how we develop and optimize algorithmic code in Mobileye REM
REM group in Mobileye is tasked with the challenge of creating and maintaining a high definition map at world scale with cm level accuracy of all road geometry and semantic elements to enable fully autonomous driving.
The map is constructed from crowd sourced anonymized data of millions of driving assistance systems running computer vision processes in consumer vehicles.
In this talk we will share stories from the trenches on how we optimized python workloads to run distributed big data processing.
We will discuss:
- how we approach these kinds of issues
- tools we use to identify and optimize algorithmic python code
- examples of how to write algorithmic code that runs 10x times faster leveraging a range of tools and technologies
Hebrew
Target audience –R&D
With 20 years of experience writing code for production across multiple environments, industries, and domains - I am an actual full stacker - be it embedded code, computer vision algorithms, Big Data engineering, or Immersive UX in mobile. But most of all I like to take hard problems and make them go away.