NJ-01 · NAJI ULLAH

BOREHOLE LOG NJ-01SHEET 1 OF 1STATUS: IN PROGRESS

DRILLED BY Naji Ullah Full-Stack Software Engineer — Lahore

SITE
LAHORE, PK
COORDINATES
31.52° N · 74.36° E
GROUND ELEV.
217 m ASL
METHOD
FULL-STACK ROTARY
COMMENCED
FEB 2025
DRILLING FOR
IN PROGRESS

This site is a drilling log of everything I’ve shipped — React and Next.js where people can see it, Django and Firebase where they can’t. Recent work lies near the surface; the foundations are at the bottom. Scroll to drill.

Begin drilling ↓Skip to contact

Experience — RECENT DEPOSITS

NJ-01/000.0–8.0 mACTIVE FILL — MOBILE / WEB / AUTOMATION DEBRISRECOVERY: 100%

SOFTWARE ENGINEER · STRUMLINES · NOV 2025 - JUL 2026

Software engineer at struMLines, where the job title said 'software' and declined to be more specific. I built mobile apps, web apps, and automations, whichever one was on fire that week. 'Cross-platform' is a lovely word. People say it the way sailors describe weather they haven't met yet.

The automations were my favourite kind of work: software whose entire ambition is that nobody ever has to think about it again. Some of it is still running somewhere in the dark, unthanked. That's the good ending.

※ Youngest deposit. Mixed composition. Holds anyway.

NJ-01/018.0–16.0 mCOMPACTED DEPOSITS — SEVENTEEN MONTHS, LOAD-BEARINGRECOVERY: 100%

SOFTWARE ENGINEER · DEVFLOVV · FEB 2025 - JUNE 2026

I was a software engineer at Devflovv. In practice that meant building responsive front ends in React and Next.js, writing the Django and DRF that sat behind them, and integrating whatever third-party API the work needed - AI services, Stripe, email - usually after reading the docs so nobody else had to.

I worked across Firebase, Firestore, Auth, Cloud Functions - set up routing and state with Redux Toolkit and Saga, and spent a fair amount of time chasing performance problems, which are almost always a redundant API call wearing a disguise. The rest is the unglamorous part: reviews, planning, and showing up.

※ Settled under pressure. As deposits do.

Selected work — seven project case studies

NJ-01/0216.0–30.0 mDENSE LAMINATED CLAYRECOVERY: 100%

Lamppost.io — FINTECH

A financial data platform that needed to stop asking its own API the same question twice.

ROLE: FRONT-END ENGINEER · ENTRY 01

I owned the school org-chart - a clear way to show who answers to whom - and then went hunting for everywhere the app double-asked the server. It was asking a lot. That kind of work is the most satisfying: nothing new on screen, everything quietly faster.

FIELD CORRECTION — Optimized API performance across the platform. Stopped asking the server the same question twice.

FIELD OBSERVATIONS

  • Async flow on Redux Saga, kept legible.
  • Built the org-chart from nothing.
  • Trimmed redundant API calls; Lighthouse calmed down.
  • Split big components into small honest ones.
  • Sat close to backend and product so things lined up.

INCLUSIONS DETECTED

  • React
  • Redux Saga
  • Performance

※ Compresses well. Fewer API calls per meter.

NJ-01/0330.0–44.0 mQUARTZ VEIN, CRYSTALLINERECOVERY: 100%

BlendGenAI — AI

Type a sentence, get a 3D model you can spin around.

ROLE: FULL-STACK ENGINEER · ENTRY 02

I built the whole front end and the viewer you inspect models in. The quietly clever part was running people's vague prompts through OpenAI first, so the generator had a fighting chance before it tried to sculpt 'cool dragon thing'.

FIELD OBSERVATIONS

  • Entire React front end.
  • Three.js model viewer on the dashboard.
  • Meshy AI for generate, enhance and texture.
  • OpenAI prompt clean-up before generation.
  • Smooth on phones, not just my laptop.

INCLUSIONS DETECTED

  • React
  • Three.js
  • Meshy AI
  • OpenAI

※ Catches light. Renders in three dimensions.

NJ-01/0444.0–58.0 mLIMESTONE, MEDICAL GRADERECOVERY: 100%

SciMedics — HEALTH

A medical SaaS where the rules are strict and the data is not allowed to be wrong.

ROLE: BACKEND ENGINEER · ENTRY 03

I lived on the backend - survey workflows, the report generation, and APIs that had to be fast and appropriately paranoid. Health software is a good teacher. It does not forgive 'probably fine'.

FIELD CORRECTION — Implemented secure healthcare data workflows. Made the database refuse to be wrong.

FIELD OBSERVATIONS

  • Survey workflows and report generation in Django.
  • APIs on DRF over PostgreSQL.
  • Access kept deliberately tight.
  • Several internal tools onto one clean path.

INCLUSIONS DETECTED

  • Next.js
  • Django
  • DRF
  • PostgreSQL

※ Calcium-bearing. Does not forgive 'probably fine'.

NJ-01/0558.0–72.0 mCONGLOMERATE — FIVE SOURCES, CEMENTEDRECOVERY: 100%

MarketSmart - Pursuit Data — DATA

Pulled business leads off five different sites and kept them searchable the moment they landed.

ROLE: FULL-STACK ENGINEER · ENTRY 04

Scrapers feeding Firestore, billing on Stripe, search on Algolia, and auth that made sure people only ever saw their own pile. Five sources is four more than anyone hopes for. It mostly held.

FIELD OBSERVATIONS

  • Next.js + Firebase, scraping five sources.
  • Firebase Extensions: Stripe, Algolia, SendGrid.
  • Role-based auth via custom claims.
  • Code reviews and sprint planning, properly.

INCLUSIONS DETECTED

  • Next.js
  • Firebase
  • Stripe
  • Algolia

※ A rock made of other rocks. Searchable on landing.

NJ-01/0672.0–86.0 mCOMPACTED FILL, TIMBER INCLUSIONSRECOVERY: 100%

Fence & Deck Connection — OPERATIONS

Quoting and workflow, glued to Salesforce, paid through Square.

ROLE: FULL-STACK ENGINEER · ENTRY 05

I kept quotes and opportunities in step with the CRM, automated the reports nobody wanted to make by hand, and wired up real money. Then I talked to the people actually using it, which changed more than the code did.

FIELD OBSERVATIONS

  • Quoting workflows in React + Firebase.
  • Quote and opportunity data synced with Salesforce.
  • Puppeteer-generated reports.
  • Cloud Functions keeping modules in sync.
  • Square ACH and card payments via Node.

INCLUSIONS DETECTED

  • React
  • Firebase
  • Salesforce
  • Square

※ Fences above, invoices below.

NJ-01/0786.0–100.0 mENGINEERED FILL, LOAD-BEARINGRECOVERY: 100%

Ryan Homes Integrations — OPERATIONS

A quote system wired into the Ryan Homes API, with dashboards that finally agreed on the numbers.

ROLE: FULL-STACK ENGINEER · ENTRY 06

Most of this was getting two systems to tell the same story. I wrote the Node logic that creates opportunities through their API, then spent a good while making the dashboards stop contradicting each other.

FIELD OBSERVATIONS

  • Quote-management workflows in React + Firebase.
  • Node logic creating opportunities via the Ryan Homes API.
  • Dashboards reconciled and trusted.
  • Steady shipping with the team.
  • Hands-on with Salesforce and CRM models.

INCLUSIONS DETECTED

  • React
  • Firebase
  • Node.js
  • Salesforce

※ Two systems, one story. Reconciled.

NJ-01/08100.0–114.0 mAQUIFER — WATER-BEARING, NO VISIBLE SURFACERECOVERY: 100%

Angi Lead Integrations — AUTOMATION

A bridge from Angi into Salesforce that took webhooks live and turned away anything malformed.

ROLE: AUTOMATION · ENTRY 07

No interface, just plumbing - multi-path lead routing on Zapier, normalised with lookup tables so it survived Salesforce's strict picklists, and locked down so the junk stayed out. Unglamorous, and very satisfying when it just runs.

FIELD CORRECTION — Developed integration middleware solutions. Plumbing. Proud of it.

FIELD OBSERVATIONS

  • Multi-path lead pipeline, Angi → Salesforce, on Zapier.
  • Real-time webhook ingestion.
  • Lookup-table normalisation for strict picklists.
  • Custom header auth on the endpoints.

INCLUSIONS DETECTED

  • Zapier
  • Salesforce
  • Webhooks
  • API

※ Invisible from ground level. It just runs.

NJ-01/LAB114.0–122.0 mSAMPLE ASSAY — COMPOSITION BY CATEGORYALL SAMPLES ANALYSED

Skills — LAB ANALYSIS

SAMPLE ASSAY · COMPOSITION BY CATEGORY · TAP A COMPOUND TO TRACE IT THROUGH THE CORE

FRONTEND

What the user actually touches.

  • JavaScript
  • TypeScript
  • HTML & CSS
  • Tailwind
  • Material UI

BACKEND

Where the rules and the data live.

  • Serverless

ON THE BENCH

Everything that helps the other two ship.

  • Git & GitHub
  • Docker
  • AWS EC2
  • Firebase Hosting
  • Postman

×N = COMPOUND DETECTED IN N OF 7 CORES. UNMARKED COMPOUNDS PRESENT IN THE TOOLBOX, NOT YET IN A LOGGED CORE.

NJ-01/BR122.0–128.0 mBEDROCK — STILL FORMINGDRILLING CONTINUES AT NIGHT

Education — BEDROCK

BS SOFTWARE ENGINEERING · COMSATS UNIVERSITY, LAHORE · 2024 - Present

Bedrock, still forming. Studying it by night, applying it by deadline. Refusal not expected before 2028.

END OF BOREHOLE — TOTAL DEPTH 128.0 m

WATER TABLE: REACHED · REFUSAL: NONE

Contact the driller

Available for interesting work. Email is the fastest route to the rig; the other two also reach it.

najiu1836@gmail.comGitHub ↗LinkedIn ↗