Yash Ladha software engineer Node.js collaborator emeritus
I make large systems faster, cheaper, and harder to break.
8+ years of system architecture, runtime internals, and distributed systems. Right now I work on UPI payments at Amazon: making them cheaper to run, easier to see into, and harder to knock over.
What happens when a dependency goes down? Break it, then add guards until requests stop hanging.
b break 1 2 3 guards- success
- n/a
- p99
- n/a
- upstream load
- n/a
- in flight
- n/a
› All green. Break the upstream and watch what callers go through.
Where I have spent my time
Hover or use the arrow keys to inspect a span.
Amazon
UPI Platform
- period
- 2021 to now
- domain
- UPI payments
- product
- Amazon Pay
- focus
- cost, observability, reliability
- Built real-time dispute resolution for UPI payments on Amazon Pay, cutting pending-transaction inquiries by 70%.
- Moved workloads to offline services that reduced infrastructure cost by 80%.
- Improved end-to-end monitoring, lowering high-severity tickets by 60%.
Open source
Projects I have contributed to, together past 10,000 stars on GitHub.
Writing
Benchmarks, tooling, and notes on how things work underneath.
- Java vs Rust Kafka Clients: Chasing Down the Gaps Benchmarking the Java Kafka client against rust-rdkafka on the same hardware, and tracking down why the first numbers showed a 16x gap. 26 min
- Lambda Performance Comparison Golang vs Java Compare performance of AWS Lambda written in Golang vs Java. 13 min
- Personal Development Environment using Docker and Neovim Discover how to set up a consistent and efficient personal development environment using Docker and Neovim. Learn to overcome environment issues and leverage containerization and a preferred code editor for a seamless coding experience. 5 min
Got a system that should be faster, cheaper, or calmer?
I like conversations about runtimes, infrastructure, and the failure modes in between. Email is the fastest way to reach me.