
01Architected the frontend
Enterprise platform for a major Irish healthcare group.
Frontend architecture + backend modules: tenant hierarchy, real-time rooms, document pipeline.
ZAIN UL ABIDIN · ISSUE 01 / 2026
Pakistan · UTC+5

The Engineer
2026 · Pakistan
Open to workSenior roles · Contract · Remote-first
See the work01Selected Work

01Architected the frontend
Enterprise platform for a major Irish healthcare group.
Frontend architecture + backend modules: tenant hierarchy, real-time rooms, document pipeline.

02Architected the full stack
Multi-tenant SaaS for generating compliance documentation.
An adaptive AI interview generating type-safe compliance documents at scale.

03Led end-to-end
4-app platform suite across the recruitment lifecycle.
End-to-end system: candidate intake, vetting, placement, billing - four apps, one backend.

04Frontend lead
Integrated travel ecosystem - app, site, admin, and CMS.
Four modules, one frontend architecture: Mapbox-powered travel app, marketing site, admin panel, blog CMS.
02Differentiator
I started on the frontend - React, TypeScript, interfaces people touch. Then I went deep where it counts: APIs, data models, queues, real-time sync. The interface is the surface. The system is the product.

Predictable props. Semantic structure. No leaky abstractions.
LCP, CLS, INP - the metrics that map to how a UI feels, not the ones that look good on a chart.
Reading order, focus, contrast, motion - layout decisions made at the start, not warnings fixed at the end.
For example
The same Zod definition produces a static TypeScript type AND the runtime check that parses untrusted JSON from an LLM, a form, or a remote API. One source of truth. A bug class removed.
// One Zod schema → TypeScript type + runtime validation.
// One source of truth. A whole class of bugs removed.
import { z } from "zod";
export const caseStudyFrontmatter = z.object({
title: z.string().min(3),
slug: z.string().regex(/^[a-z0-9-]+$/, "kebab-case"),
oneLine: z.string().min(10).max(200),
role: z.string(),
stack: z.array(z.string()).min(1),
liveUrl: z.string().url().optional(),
order: z.number().int().nonnegative(),
});
export type CaseStudyFrontmatter =
z.infer<typeof caseStudyFrontmatter>;
03How I Work
0+
Years shipping
0+
Production projects
0
Client domains
The Method
Then the piece that fits - deliberate, never cargo-cult. Every decision earns its place against what already works.
Read the patterns + anti-patterns
Compare against current best practice
Place the piece that fits - nothing more
Three Pillars
Multi-tenant data, real-time channels, distributed queues, security architecture.
Zod-validated LLM output. Adaptive interviews. Transcription pipelines.
Type-safe stack. Perceptual performance. A11y from inception. Honest scope.
05Writing
Systems engineering, applied AI, and the craft of building things that hold up in production. No fluff, no hype.
A single Zod schema gives you a static TypeScript type and a runtime validator. One source of truth. A class of bugs gone.
The best GEO tools in 2026 compared: Semrush, Ahrefs, Peec AI, Otterly, Profound, Scrunch, Goodie, and Rankscale. Real pricing, engines tracked, and my picks by budget.
A developer's playbook for getting cited by ChatGPT, Perplexity, and Gemini: content structure, schema markup, and distribution tactics that win AI-search citations.
06The Way We Work
Hiring a senior engineer should feel like hiring a partner, not a vendor. Here's how the engagement actually unfolds - from the first conversation to the team owning the system without me.
01
30 min · free
Tell me about the problem, the team, the constraint. I ask the questions an engineer would.
02
1 week
Architecture sketch, risk list, deliverables, milestones. Honest scope - no padding, no scope-creep clauses.
03
Project-dependent
Type-safe stack, perceptual perf, a11y from inception. Weekly demos. You see the same code I see.
04
Final week
Production hardening, observability, runbook. The team owns the system before I step away.
InterludeThe Mindset
A working belief.
Multi‑tenant SaaS, real‑time systems, applied AI - strip the surface and underneath I'm chasing the same thing in all of them. A system small enough to hold in one head. Honest enough to debug under pressure. Patient enough to outlive the team that built it.
Most software is too eager. It ships abstractions that didn't pay for themselves, features bolted on instead of designed in, ten lines where one good one would have done. The discipline I care about is the one that costs you - the schema deleted instead of versioned, the feature scoped out instead of in, the afternoon spent making ten lines be ten and not eleven.
That's the work. Quiet systems. Built like they're meant to stay.
04About

I trust schemas more than I trust comments. If the shape of the data is right, the code almost writes itself.
The three chapters so far
012024
Junior Software Engineer at Interactive Tech Solutions. Full-stack features in React and Node/Express, MongoDB schemas, interfaces built from Figma specs - production code from day one.
022024 → 2025
Software Engineer at Texagon. Frontend architecture for an Irish healthcare platform, a 4-app UK recruitment suite, and ComplianceBot AI - multi-tenant SaaS with Zod-validated AI.
032025 → Present
Backend engineering and applied AI. NestJS services, Postgres, queues, real-time sync - the systems underneath the interface.
Want the long version?
Read the full story07Contact
Senior product engineering - multi-tenant SaaS, real-time systems, applied AI. Always happy to talk shop, share notes, or hear about a problem you're stuck on.
The direct line