EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Hacker News·21d ago·5 min read
AI coding has made CI a bottleneck, so we reworked ours to keep up
TL;DR
Linear reworked its Continuous Integration (CI) process to address bottlenecks caused by rapid AI-driven coding. They reduced pull request wait times from over 6 minutes to just over 5 while halving runner time per test.
✦ Why It Matters
Evaluate your CI process today to identify and eliminate bottlenecks caused by increased code submissions.
Key Takeaways
- Upgrade CI infrastructure to faster third-party runners.
- Optimize job execution to minimize critical path delays.
- Reduce setup times by consolidating dependencies and caching.
- Implement efficient test execution strategies to parallelize workloads.
How It Works
Linear's CI improvements involved moving workloads to third-party runners with better performance and modernizing their toolchain. They optimized linting by rewriting rules to use static analysis instead of type information, significantly reducing memory usage and execution time.
By restructuring job dependencies and minimizing unnecessary checks, they streamlined the CI process, allowing for faster feedback on code changes.