[MM’s] Engineering Weekly #83 — Designing Systems for AI

🏗️ Modular Monolith vs Microservices: Which Should You Choose?Choosing between one deployable application and multiple services affects operating cost, delivery speed and team structure.

Compare deployment and organizational trade-offs before committing to a service model.
Match the architecture to the team’s constraints instead of following industry fashion.

Choose monoliths or microservices based on system and team constraints, not trends.
Improve cryptographic performance and compare approaches to running Java services and AI workloads.
Treat generated code as a review and verification challenge, not a replacement for engineering judgment.
Improve Git infrastructure, browser interoperability, team interfaces and project-management practices.

🧑‍💻 How 3 Java Champions are using GitHub Copilot for increased productivity How Copilot helps them solve problems and accelerate development while keeping humans accountable for outcomes.

🔐 Faster Post-Quantum Cryptography with JDK Intrinsics HotSpot intrinsics can accelerate selected Java operations through platform-specific machine code.

The JDK supports post-quantum algorithms including ML-KEM and ML-DSA.
Intrinsics can use CPU features for hashing, vector operations and modular arithmetic.

🛡️ Security Baked Into the JVM: DirtyChai scales vertically, JGDMS scales horizontally The runtime costs and operational constraints behind JVM security guarantees, including checks, audits, thread principals and grant digests.

Security guarantees can add checks and audits that affect load-test performance.
Evaluate security mechanisms against production requirements, not guarantees in isolation.

🧩 Spring AI Recipe: Rendering MCP Apps UIs in an AG-UI/CopilotKit Client Combines Spring AI, AG-UI, CopilotKit and MCP Apps to render interactive interfaces.

🧠 Spring AI Recipe: Running LLMs in the JVM with Qxotic Explore using Qxotic’s jinfer to run LLM inference inside the JVM of a Spring AI application.

🍃 Why are more and more people using Quarkus? Spring Boot vs. Quarkus How Quarkus shifts framework work toward build time to target fast startup, lower memory use and cloud-native deployment.

AI-assisted engineering
🧪 How the Claude Code team does code review — Less human review time without increasing reverts or incidents as agent-generated code grows.
🧱 Code is clay now — Agentic coding lowers the cost of changing code, shifting how developers work with it.
🧠 I’m sorry, but you still have to think — Rust and Elixir rewrites show, why benchmarks and AI-generated code still require careful evaluation.
📈 Google Optimizes Agent Development Kit’s 200K-line Python Codebase Overnight with Gemini 4 Argon — Google used a large existing codebase to study agentic optimization beyond generating greenfield applications.

Work and engineering
🏢 How employment is changing at Anthropic — How AI adoption may reshape work organization, hiring and labor demand.
🗂️ Why AI isn’t replacing project managers, it’s exposing them — AI can automate project-management tasks, while making planning quality and leadership practices more visible.
🤝 Good Interfaces Make Good Neighbors — Clear ownership, explicit priorities and a defined request process help teams resolve disagreements.

Web platform
🌉 The bridge to agent-scale Git infrastructure — GitHub describes moving nearly all repositories to Azure.
🌐 The Future of CSS — Brings browser vendors together to improve cross-browser support for web platform features.

Trending Tags