The 3 energy saving tips for your data science ecosystem
summary
Claims three concrete levers: monitor resources with AWS Cost Explorer and PSUtil to spot underused instances and schedule autoscaling or cron‑based downscaling; choose models with energy metrics (use MLflow logging, prefer EfficientNet or compare BERT vs ElMo tradeoffs) and set retrain cadence by performance drift not habit; and profile and rewrite hotspots with better data structures and algorithms to cut repeated preprocessing costs and reduce the 2.5 TWh class IT footprint.
AI-written summary · about
chapters (6)
- 0:00Energy efficiency matters
- 2:36Defining energy efficiency
- 7:35Monitor resources
- 12:21Choose models wisely
- 16:17Optimize code
- 19:48Summary
related talks
20mSustainability architecture design for ML workloads
20mBuilding GenAI That Doesn't Cost the Earth
41mWriting Greener Software Even When You Are Stuck On-Prem
23mDriving cost-effective solutions to reduce carbon footprints
19mDigital Systems & Energy Efficiency: Contradiction or Matter of Course?
07mKeynote: Finding Your Power to Accelerate to a Sustainable Future
41mSustainable Digital
34mPath to production: An environmentally sustainable approach