DPE Newsletter | Last big thing before the new year

Want to connect with Gradle? Email me at [email protected], and have a productive month!

FEATURED RELEASE

If you like Build Cache, you’re going to ♥️ Universal Cache!

Set your faces to stun: Universal Cache is here!

Your expectations from developer productivity are about to change—if you depend on keeping your builds and tests fast with Gradle build cache and Develocity Remote Build Cache, then you’ll love this revolutionary step in caching technology.

We created Universal Cache to address a core challenge facing organizations adopting GenAI. The explosion in code output, test volume, and overall build complexity has caused their CI infrastructure to become overloaded, inefficient, and expensive.

Develocity Universal Cache combines three levels of caching—made globally available and distributed using Develocity Edge—to get the greatest acceleration benefit across your entire SDLC:

Artifact Cache: Efficiently stores and restores critical build inputs such as dependencies, toolchains (like JDKs), and build system (eg. Maven, Gradle) artifacts. Downloading all of these files in one connection—versus many separate downloads over the course of the build—boosts efficiency.
Setup Cache: Provides a fast, reliable, pre-configured build environment (e.g. tasks/goals, scripts, etc.) to your entire team, ensuring every CI run is identical, fast, and stable from the moment it begins.
Build Cache: Caches and reuses tasks/goals previously executed on CI for reuse across different machines as well as caching outputs on local machines—the larger the team and the higher the build volume, the more time saved by avoiding redundant work.

Plus, with Develocity 360’s comprehensive toolchain observability already integrated, you can track Universal Cache’s impact on DORA metrics and easily see your ROI.

If you’re a Develocity customer, then just contact your Gradle representative to get started. Installation and implementation is designed to be as low-overhead as possible.

If you’re not yet a Develocity customer, you can request a trial to learn more, or simply email [email protected] to get in touch with the right people.

Learn more

BEST PRACTICES

Measure, don’t guess

Got slow builds? It’s a common instinct to solve CI bottlenecks by throwing more hardware at the problem, yet we have data that shows that 90% of CI infrastructure often sits idle, even during queuing issues. The real bottleneck usually isn't hardware, but rather process inefficiency that you can't see.

To improve DORA metrics like Lead Time and Deployment Frequency, engineers must apply the same observability standards to the "path to production" as they do to production itself. Whether it’s dependency resolution consuming 40% of build time or specific flaky tests causing the majority of failures, you cannot fix what you don't measure.

In this blog and video, Trisha Gee makes the case for a scientific approach to developer experience: stop debugging in the dark, validate your assumptions with data, and treat your build pipeline as a critical product feature.

Learn more

EXPERT TAKES

Do you even MCP?

(This image was created via MCP, so accuracy is questionable!)

If MCP is just another one of those acronyms floating around in your mind without a place to call home, it might be time for a change.

For anyone working with GenAI code, questions arise the moment you want to connect to other resources and the outside world. Luckily, the Model Context Protocol (MCP) is here. MCP acts as a "USB-C for AI," fueling a new era of developer productivity by standardizing how models connect with resources.

This article by Mohith Shrivastava, Principal Developer Advocate at Salesforce, gives a nice overview of the why, what, and how of MCP, writing:

“By standardizing the interaction between AI agents and tools, it addresses critical fragmentation issues, enhances developer productivity through intuitive ‘vibe coding’ experiences, and provides the necessary governance for enterprise-grade AI adoption.”

What makes MCP so great, in a nutshell:

Build Once, Run Anywhere: Create a single "MCP server" for a tool to make it instantly accessible across any compatible client or system.
"Vibe Coding" Reality: Move beyond complex CLI syntax—MCP lets you control tools via natural language, ensuring agents execute the correct, up-to-date commands. Note: opinions on vibe coding can be shared with this newsletter editor ;-)
Enterprise-Grade Security: Centralized gateways allow admins to allowlist specific tools and wrap actions in business policies, ensuring safe, governed AI adoption.

Incidentally, Mohith’s colleague, Venkad Dhandapani, gave a talk at DPE Summit 2025 titled “Salesforce’s AI journey for Developer Productivity: From Single Tool to Multi-Agent Development Ecosystem”, so take a look there.

P.S. Catch a glimpse of Develocity’s MCP server for troubleshooting build errors with agentic AI in this recent webinar.

Learn more

DPE UNIVERSITY

Unleash your inner Gradle build engineer!

If you’re a software developer using Gradle Build Tool, then you’ve probably got a good grasp on using it for coding, testing, and other daily tasks.

But for build engineers, who are responsible for authoring and maintaining the build, you also need to know about the internal mechanisms of Gradle to create custom logic, optimize performance, and ensure the build scales across the team.

In our newest DPE University course, Introduction to Gradle for Build Engineers, you’ll learn more about authoring custom tasks for your builds and enabling task sharing across subprojects.

Ready to empower your inner build engineer with greater knowledge and boost your professional profile?

Learn more

CAREER OPPORTUNITIES

DPE (and AI) job openings

The industry needs you! You might find your dream role among these job openings related to DPE, AI developer productivity, and engineering leadership.

NOTE: These postings are active at the time of sending but are subject to change.

Gradle Inc. | 2261 Market Street | San Francisco, CA 94114
Privacy Policy | Unsubscribe