The hottest Cost Optimization Substack posts right now

Working on pet projects is fun, but it's important to have clear learning goals to actually gain knowledge from them.
When using tools like Spark or Airflow, always ask what problem they solve to understand their value better.
To make your projects more effective, think like a user and check if they get what they need from your data systems.

A structured, reproducible forecasting pipeline models how strong human forecasters think so methods can be tested and refined systematically.
Huge cost cuts made iteration affordable: per-question cost dropped from $0.109 to $0.004 (about 27×), enabling many more experiments across the tournament.
The team accepts a likely short-term performance hit by using cheaper models and fewer tokens because the priority is learning which pipeline parts truly matter using the tournament as a feedback loop.

Security teams can optimize SIEM costs and improve data management by filtering logs effectively before they are ingested into the system. Filtering can enhance security data lake efficiency, reducing unnecessary costs and improving overall data quality.
Starting with clear intentions and asking key questions about data value, cost constraints, and threat visibility can help in creating a comprehensive and cost-efficient log filtering program.
Filtering at various stages - source, in transit, and within the SIEM itself - allows security teams to reduce storage costs, optimize performance, improve data quality, and enhance the relevance of collected logs.

In a tech company, engineering involves balancing cloud costs and user interface to optimize costs and enhance user experience.
Reducing costs significantly is crucial for a company's profitability regardless of other measures like discounts or marketing strategies.
Engineering decisions impact user experience constraints and cloud costs, requiring a balance between the two for system efficiency.

Training an early ranker to mimic the final ranker can improve top-line metrics and reduce costs
Knowledge distillation involves training a student model, the early ranker, to learn from a teacher model, the final ranker
Implementing knowledge distillation through shared or auxiliary tasks can increase alignment between the early and final rankers

Get a weekly roundup of the best Substack posts, by hacker news affinity:

Start early and consistently optimize cloud costs to control spending
Evaluate and apply tactics for optimizing compute, storage, and networking resources to reduce waste
Delegate cost optimization tasks, introduce review processes for new cloud services, and use cloud tools to support cost optimization efforts

Avoid creating Lambda monoliths to improve performance and cost efficiency.
Be cautious of Lambdas calling other Lambdas to prevent increased costs and potential bottlenecks.
Design Lambdas to be idempotent to avoid data duplication and ensure system reliability.

Using Cloudwatch Logs can be expensive due to high ingestion costs.
Consider alternative solutions like Datadog or a custom AWS setup for cost savings.
Optimizing log storage and querying processes can lead to significant cost reductions.

Check if your cloud services are optimized to keep costs down.
Separate workloads to give each part of your application exactly what it needs.
Consider delegating non-urgent tasks to save on costs with event-driven architecture.

Deploying machine learning models on GPU cores can be costly due to server prices and lack of scalability.
Using Kubernetes and KEDA for autoscaling GPU nodes can significantly reduce costs and improve scalability.
Implementing cost-optimized ML on production can be achieved by using K8s and autoscaling GPU nodes, resulting in substantial cost savings.