01 / FULL-STACK & AUTOMATION ENGINEER

MUSTAFA
FARAZ

A full-stack engineer and automation architect — building systems that eliminate inefficiency, software that ships without apology, and solutions that outlast their deadlines.

02 / ARCHITECT PROFILE

About

Mustafa Faraz — Full-Stack Software Engineer and Automation Architect based in Islamabad, Pakistan

Education

  • B.S. Computer Science

    FAST NUCES

    CGPA 3.77 — Bronze Medalist — Dean's List

    Expected Aug 2027

  • O Levels & A Levels

    The City School

    O Levels 5A*s/3As · A Levels 1A*s/2As — High Achiever

    Graduated Jun 2023

Most CS students wait to graduate before they start doing real work. I didn't.

While still in university, I landed a Software Engineer role at a US-based fintech — not an internship, an actual engineering position. I ported a Python quant library to TypeScript, built a Julia backtesting framework, and optimized file I/O using memory-mapped files. Real production work, real financial systems.

Before that, I designed and shipped a Python BDD automation framework at HBL that cut manual testing effort across releases. Then went on to build backend APIs and frontend components for a live telemedicine platform at Contour Software on the MERN stack.

On the freelance side, I maintain a 100% Job Success Score on Upwork as a Rising Talent — building automation pipelines that have turned weeks of manual enterprise work into seconds, and full-stack products for international clients.

If there's a repetitive process, I'll automate it. If there's a product to build, I'll ship it.

Technical Arsenal

Frontend

  • React
  • HTML
  • CSS
  • JavaScript
  • TypeScript

Backend

  • Node.js
  • Express.js
  • Python
  • REST APIs
  • MongoDB

Automation & Systems

  • Selenium
  • BeautifulSoup
  • Pandas
  • OpenPyXL
  • MPI
  • OpenMP
  • Behave

Languages

  • C++
  • Java
  • Kotlin
  • Julia
  • C#
  • x86 Assembly

03 / PROFESSIONAL RECORD

Experience

  1. Development & SQA Intern

    Contour Software— Islamabad, PakistanJun 2025 – Aug 2025

    Worked on a live, production telemedicine platform built on the MERN stack. Developed backend REST APIs in Node.js/Express and built frontend components in React, ensuring smooth client-server integration. Also contributed to SQA — writing test cases, reporting defects, and tracking issues in JIRA while working alongside cross-functional teams in a structured software development environment.

    • MERN
    • React
    • Node.js
    • Express
    • MongoDB
    • JIRA
    • SQA
  2. Software Engineer

    Algozen— United States, RemoteOct 2024 – Apr 2025

    Not an internship — a full engineering role at a US-based fintech startup while still in university. Engineered and maintained a Julia-based backtesting framework for financial modeling, developing new computation nodes and debugging core logic to ensure compliance with financial standards. Ported major components of the Python QuantStats library to TypeScript using Data-Forge and Lodash for time-series return analysis. Pushed system performance significantly by replacing text-based file storage with memory-mapped files, cutting I/O overhead.

    • TypeScript
    • Python
    • Julia
    • Fintech
    • Quantitative Finance
  3. Test Automation Engineer Intern

    Habib Bank Limited (HBL)— Islamabad, PakistanJul 2024 – Aug 2024

    Designed and implemented a Python-based BDD test automation framework from scratch using Selenium and Behave. Automated browser workflows that previously required manual execution across every release cycle, reducing testing effort and release time significantly. Improved test reliability and coverage across the board.

    • Python
    • Selenium
    • BDD
    • Behave
    • Test Automation

04 / SELECTED WORKS

Works

terminal
$ python alpr.py --input vehicle.jpg
→ Preprocessing image...
→ Edge detection: Canny(50, 150)
→ Contour extraction: 47 found
→ Plate region: [x:142, y:310, w:280, h:64]
→ OCR Result: ABC-1234
✓ Plate recognized in 0.84s
PYTHONOPENCVCOMPUTER VISIONDIP

ALPR – License Plate Recognition

A fully automated license plate detection and recognition system built using core Digital Image Processing techniques. The pipeline handles image preprocessing, edge detection, plate region segmentation, and character recognition — extracting license plate text from raw vehicle images without any external ML APIs. Built from the ground up using OpenCV and pure DIP concepts.

terminal
$ mpirun -np 8 ./heat2d --grid 1024x1024
→ Domain decomposition: 8 partitions
→ OpenMP threads per rank: 4
→ Iteration 1000/5000 Δ=0.00234
→ Iteration 5000/5000 Δ=0.00001
✓ Converged | Time: 2.41s
Speedup: 6.8x | Efficiency: 85%
C++MPIOPENMPPARALLEL COMPUTING

Parallel Heat Equation Solver

A high-performance parallel numerical solver for the 2D heat equation, implementing domain decomposition across multiple processes using MPI and shared-memory parallelism using OpenMP. Benchmarked across multiple thread/process configurations to analyze speedup, efficiency, and scalability. Demonstrates a solid command of low-level parallel systems programming.

terminal
$ python los_automation.py
→ Loading raw field data... 2,847 rows
→ Step 1/6: Data ingestion ✓
→ Step 2/6: Validation ✓
→ Step 3/6: Transformation ✓
→ Step 4/6: Computation ✓
→ Step 5/6: Formatting ✓
→ Step 6/6: Excel export ✓
✓ Pipeline complete: 3.2s (was: ~1 month)
PYTHONPANDASOPENPYXLAUTOMATIONENTERPRISE

Excel LOS Automation

Built for a real oil and gas company client. Their team received raw field data files and had to manually process them through a 6-step workflow — a month-long cycle, every cycle. I automated the entire pipeline end-to-end: data ingestion, transformation, validation, computation, and formatted Excel output. What used to take a month now runs in seconds. Delivered as a production tool actively used by the client.

socially_–_instagram_clone.code
class SocialApp : Application() {
// Firebase + WebRTC Integration
val auth = FirebaseAuth.getInstance()
val rtc = PeerConnectionFactory()
 
fun initVideoCall(peerId: String) {
val offer = createOffer(constraints)
signalingChannel.send(offer)
}
}
KOTLINANDROIDFIREBASEWEBRTC

Socially – Instagram Clone

A full-featured Instagram clone built natively in Kotlin for Android. Not a simplified tutorial project — this includes a complete friend system, real-time messaging, audio and video calling, post creation and sharing, stories, profile pages, and a feed algorithm. Built with Firebase for backend services and WebRTC for peer-to-peer audio/video communication. A full social networking application from scratch.

05 / TECHNICAL STACK

Skills

Languages

PythonJavaScriptTypeScriptC++JavaKotlinJuliaC#x86 AssemblyHTMLCSS

Frameworks & Libraries

ReactNode.jsExpress.jsMERN StackSeleniumPandasNumPyBeautifulSoupBehaveData-ForgeLodashOpenPyXLSFMLWebRTC

Domains

Full-Stack DevelopmentTest AutomationWeb ScrapingREST APIsBDDParallel ComputingComputer VisionAndroid DevelopmentBackend EngineeringQuantitative Finance

Tools

GitJIRAVS CodePyCharmAndroid StudioVisual StudioPostmanFirebase

06 / INITIATE CONNECTION

Contact

Let's Build
Something Real.

Open to full-time roles, internships, and freelance projects.

LOCATION

Islamabad, Pakistan. Available for remote work worldwide.

SOCIALS