Clever had two critical and complex projects to serve 940+ school districts with 12.6M daily API requests on their K-12 platform: automated data processing + self-service identity management. All on a tight deadline so school district administrators could be ready ahead of the school year starting.

Your coding agent will write tests. What it won't do, by default, is write good tests. Mutation testing fills that gap, and it turns out LLMs benefit from it far more than human developers. Neal Lindsay breaks down the set up and getting your agent to run it. https://link.testdouble.com/c8f60d

AI doesn't have an accuracy problem. It has a target problem. Three Amigos, adapted for solo devs: BDD → TDD → Given/When/Then Acceptance Criteria. Stop correcting after the fact. Approach and useful framework tools: https://link.testdouble.com/bcf238

Staff Consultant Brian Hughes breaks down the myth that software tests slow down developers. If you're doing it right, software tests can actually speed you up. youtu.be/n1792fDDxGA?...

We're here for you when you need a pair you can trust. We take a pragmatic approach to using AI, and can help you get the most out of it. Book a 1:1 office hours session: link.testdouble.com/d28109

Verse Communications was building decentralized social infrastructure, and needed a repeatable path to test assumptions and find product-market fit fast. We ran rapid AI prototyping sprints and built discovery frameworks so they could move with confidence. https://link.testdouble.com/a004f1

Welcome back to Jackie Kalb, who has returned as our Senior Client Partner!

Standard Ruby (the bikeshed-proof linter & formatter) is used in over 9,000 projects. But how does the team at Test Double decide which RuboCop rules make the cut? Camilo Payan shares the process: the criteria, the tradeoffs, and what they deliberately leave out. https://link.testdouble.com/dde93d

Some problems don't need AI Ralph loops. They need someone with experience to cut through the noise. Our free pairing sessions connect you with practitioners who've spent 15+ years developing judgment to know what matters, and what to ignore. Book office hours at link.testdouble.com/d28109

Staff Consultant Steve Jackson talks about why the counterintuitive point of testing is confidence that something will fail ... so you know what messed up. Why is that important? It actually increases your speed, because you know tests will alert you if something's not right. youtu.be/yumiz1-54k8

You can generate more code faster than ever. But speed was never the goal—it's a byproduct of achieving outcomes under constraints. AI amplifies quality and entropy equally. The key is the harness you build around it. Read the post and let us know what you think https://link.testdouble.com/17569f

Product management transformation: From chaos to 500% growth Before: $6.5B retailer, talented teams, 12 initiatives, missed deadlines, 14 engineers quit in 6 months, 40% funding cut After: 5 focused initiatives, clear strategy, data-driven decisions, 1x/quarter → 2x/month releases (500% increase)

This approach to unstructured video calls might change how you build relationships with remote teams. Counterintuitively, it's about embracing the awkward silence until it becomes comfortable. https://link.testdouble.com/220119

Misalignment between team structure and product needs is one of the root causes of low value delivery. Real agility is structural. Test Double VP of Delivery Doc Norton shares thoughts. https://docondev.com/blog/2025/4/7/composition-isnt-just-about-code

Staff Consultant Steve Jackson shares why we have a pragmatic mindset around failing software tests. www.youtube.com/watch?v=WU8R...

We use AI to solve business problems, not for hype's sake. Solving problems with software to drive business impact is why we exist. CEO Todd Kaufman shares why it's important for business leaders to periodically revisit organizational mission, vision, and values. https://link.testdouble.com/6a0803

Phil Dunphy, the dad from Modern Family, would tell his family, "slow is smooth and smooth is fast," often when the family was panicking and neither smooth nor fast. You should read @docondev.bsky.social's exploration of the same idea for software dev teams BEFORE your team is panicking. 😉

79% of healthcare tech leaders cite limited budget as their top blocker. Yet the real constraint isn't money—it's solving the wrong problems. Our ongoing research points to patterns. And we've got a useful framework to start diagnosing root causes. https://link.testdouble.com/389f5e

Microservices scaling: MedTech with 50+ services, 235k daily requests Mission: Streamline and balance speed with quality Fix: Strategic simplification Results: 50% faster deployments, resolutions in minutes Embedded advisory solving problems alongside your team > consultant recs that disappear

Enbala needed to scale an energy management platform to process gigabytes of data per second from thousands of grid assets in real time. We embedded for rapid feature releases, performance refactoring, automated testing, and mentoring devs and product managers. https://link.testdouble.com/e280de

When leaders demand speed, teams often cut corners—creating brittleness that slows everything down. @docondev.bsky.social has advice: Real speed emerges from flow, small batches, and systems designed for adaptation, not pushing harder on execution. https://link.testdouble.com/0655f3

Rails upgrades in massive production codebases: a case study from Gusto. Challenge: Upgrade legacy Rails version without pulling experts from critical feature work

There's a potential disconnect in healthcare on AI spend vs. resources. We surveyed technical and product leaders: 84% increased AI spend while 79% cite limited resources as top barrier. More insights in our report, plus survey is open to continue gathering data! https://link.testdouble.com/389f5e

You shouldn't have to choose: clean Python code or poorly-named external APIs. Pydantic field aliases let you have both: Pythonic variable names in your code, whatever naming scheme the API expects on the wire. No more manual field mapping or workarounds. https://link.testdouble.com/ddffe7 #Python

Stuck on a tough challenge? ➡️ Debugging a gnarly codebase? ➡️ Wondering whether to do a rewrite or refactor your legacy system? ➡️ Evaluating if your product idea is solving the right problem? We've got you covered: free, no strings attached office hours Book here 👉 https://link.testdouble.com/x5r

Your velocity problem isn't a headcount problem. It's an unclear requirements problem. A technical debt problem. A tooling problem. Hiring more developers to work around broken processes just amplifies the chaos. https://link.testdouble.com/904ed5

Principal Software Consultant Dave Mosher talks about identifying the legacy modernization strategy that's right for you and your codebase. Sometimes a seams path makes sense, and in other cases following a value path is better suited. youtu.be/CMqfSFIUiDU

You need to fight for feedback. We often front-load discovery to feel confident before building—but that's not real agility. Continuous learning reduces risk better than perfect plans ever could. https://link.testdouble.com/5b3945

"I was elated when another engineer mentioned deploying a new dashboard without my involvement." We don't focus only on the technical solution—we also build team capabilities. Our approach to observability, and why the human element matters most: https://link.testdouble.com/a882ec

When a wide range of things aren't working or didn't scale effectively, you hit trade off decisions. To make an effective business case for budgeting you need to ensure you're advocating for the right things. Our pragmatic and actionable assessments help you do just that.

The critical things that ensure your software application is maintainable now and in the future are seemingly simple. Brian Hughes shares two key truths about automated tests and legacy software applications. youtu.be/VGIL5BN6RzY

Slow delivery isn't usually about your people—it's about your system. Shifting focus to incremental improvements in the system helps change not just processes but behaviors for lasting change. https://link.testdouble.com/09fc66

Healthcare tech teams keep launching features. Tech debt keeps growing. The problem isn't your developers—it's solving symptoms instead of root causes. This report points to possible reasons why. https://link.testdouble.com/389f5e

If you rely on Ingress NGINX pay attention! https://kubernetes.io/blog/2026/01/29/ingress-nginx-statement/

It takes many years of lessons learned to scale software without scaling problems. We've made every mistake so you don't have to! Find out how we helped an early-stage startup make iterative improvements to increase efficiency and the bottom line as they scale. https://link.testdouble.com/076e97

Staff Consultant Steve Jackson shares when a legacy software app rewrite makes sense vs a refactor approach. youtu.be/f1u-8YYm4Og

Learn advanced Pydantic techniques to bring order to chaotic, real world data. In this talk @kadams54.bsky.social gave for IndyPy he introduces Pydantic for Python with practical examples. https://link.testdouble.com/1fc497

In healthcare, challenges around interoperability, compliance, and data abound. A pragmatic approach that's focused on driving the business forward helps cut through the noise. Many thanks to Kiley Blake for contributing insights to our report, Symptoms to Sources!

Learn how to validate datatypes that go beyond Python's primitives. This blog from Kyle Adams & Gabriel Côté-Carrier covers different validators included with Pydantic as well as how to write your own custom validators. And it's post 4 in an ongoing series! https://link.testdouble.com/09072d

Many teams get stuck choosing between massive rewrites (unrealistic) or good intentions (unsustainable). There's a better way. An incremental, ratcheting approach can give you more leverage for lasting impact. We offer assessments that can help you find the seams and map out a plan to get started.

The problems most worth solving are often the slippery ones. The problems that are hard to define the edges of. Many thanks to Mike Berkman for contributing insights to our report on healthcare challenges, Symptoms to Sources!

Python type safety lightbulb moment: mypy = compile-time safety Pydantic = runtime safety Helped a client shift domain knowledge from heads → codebase with this combo. https://link.testdouble.com/769eb9 No more "wait, what does this field contain?" moments 🎯 #python #pydantic

Staff Consultant Brian Hughes shares why a seams approach to legacy modernization can be even more complicated when seams are crossed. This introduces multiple possible paths. Luckily for you, we've encountered that before 💚 www.youtube.com/watch?v=Ilk7...

50+ microservices. Lean team. Daily fires. The fix: - Custom logging library - API gateway architecture - Structured learning via katas - Improved Collaboration 50% faster deployments, resolutions in minutes. Software problems are human problems. We speak human. https://link.testdouble.com/1bd181

Product transformations fail when you try to change behaviors without naming them first. Not taglines. Real principles that constrain decisions. Trade-offs that acknowledge tension. Example: Quality even over speed. Without this autonomy becomes chaos. https://link.testdouble.com/e34baf

Increasingly software problems are data problems, and this is especially true in healthcare. Many thanks to Andrew Warner for contributing to our report, Symptoms to Sources!

Complex systems are usually complex for incidental reasons. It's counter-intuitive, but purposefully simplifying can be huge for scaling.

When you're weighing a legacy rewrite vs refactor, consider what stays and what doesn't. You don't need everything in a new or updated system. Staff Consultant Steve Jackson says cutting scope can align with a rewrite, and deserves review with your product team. www.youtube.com/watch?v=4N4y...

Code is a means, not an end. The code itself is ephemeral. What endures is the solved problem, the working system, the user who can now do their job better. That's why, while we use AI to solve interesting challenges, we say software problems are human problems.

In healthcare there are so many areas you could choose to solve problems and improve inefficiencies. When you focus on the voice of the customer and improving patient outcomes, that's when you get to bigger impact. Many thanks to Sara Saldoff for sharing insights for our report, Symptoms to Sources.