Code & Curiosity: Where I Am Now

Reorienting Focus

In software engineering, resting on your current knowledge is a quick way to fall behind. Lately, my focus has shifted heavily towards understanding and utilizing AI-driven development workflows, while continuously refining my core competencies in React and TypeScript. In this context, building applications is no longer just about writing code; it is about leveraging advanced tools to design more robust systems.

Building at Scale: The Company Web Framework

Currently, my main professional focus is developing a company-wide web framework library. The primary goal of this library is to standardize frontend components across teams, ensuring consistency while maintaining strict requirements for performance and speed.

Building a shared framework is significantly different from developing standard client applications. Every component must be highly reusable, type-safe, and compile cleanly under various bundlers. Because different teams use different rendering environments, ensuring compatibility has required deep-diving into runtime performance and optimizing rendering cycles.

Home Labs and Server-Side Rendering (SSR)

While my professional work centers on standardizing UI components, my personal time is dedicated to exploring Server-Side Rendering (SSR) architectures, specifically with React Server Components (RSC) and NextJS.

Developing for SSR introduces distinct challenges. You must carefully design component boundaries, separating static markup from dynamic client-side logic. At home, I spend a lot of time crafting components that load instantly without triggering hydration errors. By separating concerns and utilizing NextJS's server-driven layout model, I can achieve fast page transitions and clean data loading states.

The Developer Ecosystem: Claude Code and Antigravity

A key driver of my progress today is the integration of agentic AI coding assistants. I employ Claude Code to accelerate professional projects at work, and Antigravity to power my personal experiments at home.

These tools are not just autocomplete engines; they act as pair programmers. By gathering context from historical commits, codebase structures, and team collaborative data, I can feed this context back into the LLM system. This feedback loop refines the prompt instructions and improves prompt context, enabling the AI engine to generate highly contextual, correct solutions. Balancing Claude Code and Antigravity has significantly optimized my productivity, allowing me to focus on architectural decisions rather than routine boilerplate.