DPE Summit 2025 Gradle: Why this is one of the most rewarding conferences about engineering productivity (Category Productivity)
I like the conference. DPE (Developer Productivity Engineering) Gradle is held annually in San Francisco. I've often told you about the summit. 2024 year (For example, here: 1, 2, 3 and 4). This event focuses on engineering productivity and DX. (developer experience) And this year it's gone. 23–24 In September, the organizers posted the recordings of the sessions. (here playlist).
Why I think this event is important and interesting is because DPE is not about looking at development productivity in the forehead of how many lines of code an engineer wrote, but rather as an engineering challenge: How to speed up feedback loops and reduce cognitive loag How to improve CI/CD (Faster, simplify, make cheaper) How to improve observability and simplify troubleshooting Gradle and dpe.org explicitly describe DPE as a practice that enhances developer productivity and happiness through tools, automation and data, not just management metrics.
The most valuable thing about this conference is not abstract talk about “developer productivity”, but cases from large engineering organizations. (||where water might still be||). The program included speakers and reports from Netflix, Google, Meta, Microsoft, Capital One, Uber, JetBrains, Okta, Lloyds Bank, Square, Databricks and other companies. They discussed, among other topics, the impact of GenAI on delivery, approaches to measuring the effect of AI on productivity, build acceleration, universal cache, merge queue, flaky tests, dependency updates, observability and DevEx metrics.
In general, the main conclusion of the summit is that GenAI speeds up code writing, but does not guarantee an acceleration in the delivery of working software. In keynote, Hans Docter directly formulates the question:
You can write code faster. Can you deliver it faster? The logic of the question is simple: the more code AI generates, the greater the load on build, test, compliance and deployment. If the pipeline isn’t ready, then AI doesn’t speed delivery, it just moves the bottleneck down the chain. Hence the focus on smarter pipelines, observability, predictive test orchestration, caching and shift-left practices.
I've only looked at a few reports so far, but it's clear that not all keynote reports are abstract. Max Kanat-Alexander of Capital OneWhat Makes a Great Developer Experience?The three things that improve DevEx are faster iteration, more focus time and less cognitive load. Netflix takes apart productivity in the GenAI era and separately shows how they approach build vs buy and AI coding agents implementation. Google is discussing how to actually improve the impact of AI on developer productivity. Uber shows the merge queue theme at scale, and Square shows how to use the idle time of development machines to speed up workflows and improve cache hit rate.
The value for tech executives is that the DPE Summit shows that productivity is no longer an HR discussion or a velocity debate, but a full-fledged field of platform engineering and engineering management. Measure not only output, but also flow, friction, Pipeline reliability, feedback time, cost of changes and real developer experience. And if you look at the reports, you can move from superficial conversations in the style of “AI will replace developers” to discussing changes in the entire delivery contour – from the local environment and tests to pipeline, observability and managed DevEx investments.
In the coming days, I will be talking a lot about interesting reports from this summit.
#DevEx #Metrics #DevOps #Engineering #Software #Management #Leadership