FinOps is an operational framework and cultural practice that brings together technology, finance, and business teams to collaborate on data-driven spending decisions. It's not just a tool or a dashboard - it's about shared responsibility for cloud costs across the entire organization.
For engineers, this means: You own the cost of what you build. Just like you're responsible for code quality and security, you're responsible for cost efficiency.
Click each pillar to learn more
For your capstone: Tag your resources by feature/component so you can see which parts cost the most.
For your capstone: Start with smaller instances and scale up only if needed.
For your capstone: Add a cost section to your project README.
Everyone. FinOps is a team sport.
Same training workload, 95% cost reduction
ML accelerators are 10-50x more expensive than CPUs. Use them only when needed, and use spot instances for training when possible.
Training is bursty and can tolerate interruption. Inference needs to be always-on. Design accordingly.
Training datasets can be huge. Store raw data in cheap cold storage, processed data in faster tiers.
Large models (GBs each) add up. Keep only the versions you need, archive or delete old experiments.
1. What does FinOps mean?
2. Which is the best candidate for spot instances?
3. How can you reduce storage costs for old logs?