Notes on how weactually build.

Engineering judgement written out in full — architecture decisions, framework trade-offs, and the profiling-first approach behind the performance work. Not a content calendar, and nothing here is written to fill a schedule.

Posts: 3
Written by: Bhavesh Bhoi
Topics: Architecture, performance
Cadence: When there's something worth saying

Engineering

How I Structure Next.js Applications for Long-Lived Products

The boundaries I draw before writing a feature: where the data model lives, how routes stay thin, and why the structure that feels slow in week one is what keeps week fifty cheap.

10 September 20263 min read

Architecture

React vs Next.js: Choosing the Right Tool for a Client Project

"Just use Next.js" is right most of the time and wrong often enough to be worth a real answer. How I actually decide, project by project.

22 August 20262 min read

Performance

What Actually Moves Core Web Vitals in Production

Lighthouse scores and real field data disagree more often than teams expect. A profiling-first approach to LCP, INP and CLS, instead of a best-practices checklist.

30 July 20263 min read

Working on something similar?

These are the samedecisions we make for clients.

Every post here is a judgement call we've actually made on a real build. If you're weighing the same kind of decision, it's worth a conversation before the architecture is set.

Start a project See what we do

How I Structure Next.js Applications for Long-Lived ProductsHow I Structure Next.js Applications for Long-Lived Products

React vs Next.js: Choosing the Right Tool for a Client ProjectReact vs Next.js: Choosing the Right Tool for a Client Project

What Actually Moves Core Web Vitals in ProductionWhat Actually Moves Core Web Vitals in Production

These are the samedecisions we make for clients.

How I Structure Next.js Applications for Long-Lived Products

React vs Next.js: Choosing the Right Tool for a Client Project

What Actually Moves Core Web Vitals in Production