Sonam Gupta and Kunal Nawale, SigLens
Optimizing performance isn’t about guessing — it’s about measuring. Yet, many teams skip systematic profiling and jump straight into rewriting code or scaling infrastructure. The result? Wasted time, minimal gains, and recurring issues.
In this talk, we’ll break down five high-impact tips to make profiling a practical, results-driven part of your performance workflow. You’ll learn:
- How to choose the right profiler for your language and environment.
- Where performance bottlenecks really hide (hint: it’s not always your code).
- What to look for in flame graphs, CPU/memory profiles, and I/O traces.
- How to set up lightweight, continuous profiling without slowing down production.
- Ways to prioritize performance fixes that actually improve user experience.
Whether you’re debugging a slow backend or shaving milliseconds off a critical path, these techniques will help you go from vague complaints to concrete wins — backed by data, not guesswork.
Sonam Gupta is a Founding Software Engineer at SigLens, where she focuses on building high-performance observability solutions that are both fast and cost-effective. With a passion for distributed systems and developer experience, she is dedicated to making observability accessible to all.
Kunal Nawale is a Founder/CEO of SigScalr. Previously he worked as Observability Architect at Salesforce responsible for providing monitoring solutions to 10,000+ salesforce engineers. Kunal has spent 20+ years in software development. He has been mentoring young engineers for the past 12+ years. Kunal did his Bachelors from University of Pune, and M.S. from UMass Lowell, USA.
author = {Sonam Gupta and Kunal Nawale},
title = {Profiling Your Code: 5 Tips to Significantly Boost Performance},
year = {2025},
address = {Dublin},
publisher = {USENIX Association},
month = oct
}