Lithography drives semiconductor world and out moderns way of life. Noone can imagine his/hers life without: computer, smartphone, YouTube, smartwatch, etc. At the same time everyone expects that every year all those electronics miracles will become more powerful, bring even more features, get smaller (ideally invisible). ASML lithography scanners (Twinscans) are used to produce 85% of all chips (CPUs, RAM, SSDs, NAND, bluetooth, sensors, etc) made on silicon.
ASML scanners have multiple computers inside with multiple operating systems, running software stack of >50 mil lines of code on multiple languages from C till Python and Julia. Those scanners are able to position silicon wafers with nanometer accuracy and several G of acceleration. To create, improve, and maintain the huge sw stack ASML has >1000 engineers, spread across multiple continents, timezones, cultural backgrounds.
This talk will give insights on how sw integration is organized/implemented at ASML:
- Automated developer delivery pipeline to main archive
- Every day new version is created and tested (fully automatically)
- Late bug detection (after integration) and automatic removal of bad deliveries
- Utilizing power of Google cloud for testing
- Utilizing power of Microsoft Azure/ InfluxDB/Grafana to enable 24/7 monitoring with alerts for "non self recoverable" disasters
- Wide landscape of multiple services that all have to work together to enable 24/7 integration/ build/install/testing/test results analysis
Also interesting (and quite unique) aspect of : "Lithography scanner lifetime >30 years, and customer wants to get new sw features/have support" will be explored. Not many companies in the whole world have such constrains.
Professionally deformed / pragmatic / full stack / hardcore / data driven / "hands on" / senior principle arch with not only solid all around tech background, but also killer presentations skills, ability to explain complex issues in simple terms even to board of management, and 120% can-do attitude!
From Basic and Sinclair 1982 to Cloud and K8S. Any languages, any platforms. Team sizes from 1 to 100s.
• Deflate effort estimates
• Upgrading huge legacy software stacks
• Inspiring presentations
• Forming cross department teams