Author: Gabo G.
-
Modular Monoliths Resurge in Cloud-Native Architectures
The Weekly Radar The Context In the past week, there’s been a noticeable shift toward modular monoliths among teams aiming to balance development velocity and operational simplicity. Unlike pure monoliths, a modular monolith enforces clear module boundaries within a single deployment unit. This approach preserves the local performance benefits of co-located code while allowing separate…
-
Building the AI Engineering Stack: Spec-Driven Development Takes Center Stage
The Weekly Radar The Context AI coding agents have evolved from simple autocomplete to complex ecosystems capable of planning features, modifying entire repositories, executing tests, and coordinating multi-agent workflows. In response, a set of disciplines—Spec-Driven Development, context engineering, specification engineering, agentic engineering, and harness engineering—has emerged to tame these autonomous tools. A “harness” determines what…
-
When AI-Assisted Code Backfires: Lessons from a Slower Optimization
The Weekly Radar The Context A recent Medium post titled “His First Optimization Made It Slower” recounts how a team using AI-generated code saw throughput gains and rising confidence in their pipeline metrics. Two days after deployment they applied a manual tweak to inline a critical loop only to discover overall latency had increased. The…
-
Why Monolith-First Is Making a Comeback in 2026
The Weekly Radar The Context The monolith-first approach is gaining traction as teams wrestle with the operational overhead of managing dozens or hundreds of microservices. Instead of jumping straight into a distributed architecture, squads are opting to start with a single, modular codebase. This lets them ship features more quickly, onboard engineers without steep service-mesh…
-
Mandated AI Adoption Backfires: When Speed Costs Explainability
The Weekly Radar The Context A recent Medium case study titled “Our CEO Wanted Every Developer Using AI. Six Months Later, Nobody Could Explain the Codebase.” describes how a well-intentioned mandate to integrate AI tools into everyday workflows delivered on speed—backlog clearance and slashed implementation times—but introduced a hidden cost. Over time the team lost…
-
Building Enduring Engineering Leverage in the AI Era
The Weekly Radar The Context Last week at LeadDev New York, Kaya Thomas laid out a roadmap for engineering growth in the age of AI. She argues that human-centered engineering, systems thinking, and data modeling remain the pillars where engineers can outpace code-generating tools. Rather than chasing syntax, teams must design for empathy, anticipate boundary…
-
When Agents Take the Wheel: AI Refactors 300K Lines in Three Weeks
The Weekly Radar The Context InfoQ’s case study “Agents Refactor 300K Lines in Three Weeks” chronicles a team deploying agentic refactoring, where LLM-based agents autonomously break down and apply code changes across a large codebase. Over a span of three weeks, these agents parsed, analyzed and restructured 300 000 lines of legacy code, coordinating tests…
-
Leading vs Managing: Bridging Two Pillars of Engineering Success
The Weekly Radar The Context In engineering organizations, leadership and management aren’t competing roles but complementary functions that drive team performance. Leadership is about setting a future-oriented vision, inspiring innovation, and navigating strategic trade-offs. Management focuses on day-to-day operations—triaging tasks, maintaining quality control, and developing individual skills. Effective teams blend both to achieve high velocity…
-
Service Mesh and Orchestration: The Next Step in Cloud-Native Adoption
The Weekly Radar The Context Cloud-native architectures promise scalability and resilience by decomposing applications into loosely coupled services. Yet as organizations embrace containers and microservices, managing the interactions between dozens or hundreds of services becomes a new challenge. Service orchestration patterns—centralized control planes that coordinate workflows—and service meshes—sidecar proxies that handle communication, security, and observability—are…
-
Agentic Development Lifecycles: When AI Becomes Your Teammate
The Weekly Radar The Context Traditional software development follows a series of handoffs—Requirements → Design → Build → Test → Release → Monitor—where humans shuttle context between stages. The emerging Agentic Development Lifecycle (ADLC) reframes this into artifact-driven loops: Intent → Design → Build → Evaluate → Govern → Deploy → Observe → New Intent.…