Mercedes Bernard, Staff Software Engineer at Kit, joins Robby to talk about what it really means to write code that lasts—and who it should be written for.
In this episode of Maintainable, Mercedes shares a thoughtful and practical perspective on working with legacy codebases, managing technical debt, and creating a team culture that values maintainability without fear or shame. Her guiding principle? Well-maintained software is friendly software—code that is understandable and approachable, especially for early-career developers.
Together, they discuss how to audit and stabilize older systems, avoid full rewrites, and create consistent developer experiences in large applications. Mercedes reflects on her decade in consulting and how that shaped her approach to navigating incomplete documentation, missing historical context, and multiple competing patterns in a codebase. She breaks down different types of technical debt, explains why not all of it is inherently bad, and offers strategies for advocating for maintenance work across engineering and product teams.
The conversation also touches on architecture patterns like job fan-out, measuring performance regressions, reducing infrastructure load, and building momentum for improvements even when leadership isn’t actively prioritizing them.
If you’ve ever felt overwhelmed by a messy project or struggled to justify maintenance work, this episode will leave you with a fresh mindset—and a few practical tactics—for making code more sustainable and inclusive.
Episode Highlights
[00:01:08] Defining Well-Maintained Software
Mercedes explains her top metric: software that feels friendly, especially to early-career developers navigating the codebase for the first time.
[00:03:00] What Friendly Code Actually Looks Like
She shares why consistency, discoverability, and light documentation (like class comments or UML snippets) can make a huge difference.
[00:05:00] Assessing Code Like a House Tour
Mercedes introduces her metaphor of giving a house tour to evaluate code: does everything feel like it’s in the right place—or is the stove in the cabinet?
[00:06:53] Consulting Mindset: Being a Guest in the Codebase
With a decade of consulting experience, Mercedes shares how she navigates legacy systems when historical context is long gone.
[00:10:40] Stabilizing a Startup’s Tangled Architecture
She walks through an in-depth case study where she helped a client with multiple abandoned services get back to stability—without a rewrite.
[00:17:00] The Power of a One-Line Fix
Mercedes shares how a missing check caused a job to fan out 30 million no-op background jobs a day—and how one line of code reduced that by 75%.
[00:23:40] Why State Checks Belong Everywhere
She explains how defense-in-depth patterns help avoid job queue flooding and protect system resources early in the fan-out process.
[00:24:59] Reframing Technical Debt
Not all debt is bad. Mercedes outlines three types—intentional, evolutionary, and time-based—and how to approach each one differently.
[00:28:00] Why Teams Fall Behind Without Realizing It
Mercedes and Robby talk about communication gaps between engineers and product stakeholders—and why it’s not always clear when tech debt starts piling up.
[00:34:00] Quantifying Developer Friction
Mercedes recommends expressing technical debt in terms of lost time, slow features, and increased cost rather than vague frustrations.
[00:42:00] Getting Momentum Without Permission
Her advice to individual contributors: start small. Break down your frustrations into bite-sized RFCs or tickets and show the impact.
[00:45:40] Letting the Team Drive Standards
Mercedes encourages team-led conventions over top-down declarations, and explains why having any decision is better than indecision.
[00:47:54] Recommended Reading
She shares a surprising favorite: The Secret Life of Groceries, a systems-thinking deep dive into the grocery industry by Benjamin Lorr.
Resources & Links
- 🧠 Mercedes Bernard’s website
- 🛠 Kit
- 📘 The Secret Life of Groceries by Benjamin Lorr
- 🔎 Fan-out Pattern – Martin Fowler (for deeper context on fan-out job strategies)
- 💬 Mercedes on LinkedIn
Thanks to Our Sponsor!
Turn hours of debugging into just minutes! AppSignal is a performance monitoring and error-tracking tool designed for Ruby, Elixir, Python, Node.js, Javascript, and other frameworks.
It offers six powerful features with one simple interface, providing developers with real-time insights into the performance and health of web applications.
Keep your coding cool and error-free, one line at a time!
Use the code maintainable to get a 10% discount for your first year. Check them out!
