Hello, I'm
Zain Ullah
Building intelligent applications at the intersection of data science and modern web development. 4+ years turning complex problems into elegant, production-ready solutions.
const developer = {
name: "Zain Ullah",
title: "Data Scientist & Full-Stack Dev",
education: "MS Data Science, Sapienza ",
skills: [
"React/Next.js", "Python/ML",
"TypeScript", "NLP",
],
passion: "Building intelligent apps",
hireable: true,
};
About Me
Get to know who I am and what drives me
I'm a Data Scientist & Full-Stack Developer with a Master's in Data Science from Sapienza University of Rome. My work lies at the intersection of data-driven intelligence and production-quality web engineering.
With 4+ years of combined experience, I've built everything from NLP pipelines achieving 80% classification accuracy on insurance call-center data to full-stack SaaS platforms with AI-powered features, visual workflow editors, and usage-based billing systems.
I thrive at the full product lifecycle — from designing scalable architectures with Prisma, PostgreSQL, and Next.js server actions, to training deep learning models for computer vision and NLP tasks. Published researcher with work in face recognition analysis (ASTRJ 2021).
Location
Quetta, Pakistan
Education
MS Data Science, Sapienza, Italy
Languages
English, Urdu, Italian(Basic)
Skills & Technologies
Technologies I work with across the full stack and data science spectrum
Core Proficiencies
Featured Projects
A selection of projects showcasing my range across full-stack development and data science
ENT healthcare clinic
Healthcare platform with patient self-service, appointment scheduling, and admin dashboard.
WAR Activities Detection
YOLOv3-based object detection and behavior classification system for video analysis with 91% improvement.
Brain Tumor Classification
Medical imaging classification using CNN and zoning technique achieving 97% accuracy.
All Service Provider
Android recommender system for home services using matrix factorization with Django backend.
Buzzly
Responsive social media app with modern auth, state management, and real-time data fetching.
Smart Door — Face Recognition Access Control
AI-powered smart door system using ESP32-CAM, Raspberry Pi, FaceNet, and MTCNN with liveness detection.
Professional Experience
My career journey across data science and full-stack development
Full-Stack Developer
Freelance & Personal Projects
Remote
- ›Led full product lifecycle of NexScrape, a web automation SaaS with visual workflow editor (React Flow), Next.js backend, Prisma + PostgreSQL, Clerk auth, and Stripe billing.
- ›Integrated OpenAI API for AI-powered data extraction, converting unstructured web content into structured JSON with 95%+ accuracy.
- ›Built UXPro Solutions — a premium digital agency website featuring Three.js 3D visuals, Framer Motion animations, and glassmorphism-driven UI for a software house brand.
- ›Developed Nutrium Pharma — a pharmaceutical company platform with product catalog, clinical compliance branding, and login portal for healthcare professionals.
- ›Shipped ENT Care Clinic healthcare platform with Appwrite, patient self-service, and admin dashboard, reducing manual scheduling by 30%.
- ›Built NexQit — an interactive software company website with Three.js WebGL rendering, Framer Motion micro-interactions, and SEO/performance optimization.
Data Scientist
FASI - Fondo Assistenza Sanitaria Integrativa
Rome, Italy
- ›Built sentiment analysis pipelines using Hugging Face transformers on customer feedback data, achieving 80% classification accuracy for topic extraction and satisfaction scoring across call-center interactions.
- ›Conducted annual root-cause analysis on insurance claims data to identify recurring instability factors and loss drivers, delivering actionable reports that informed strategic decision-making.
- ›Automated repetitive data processing and reporting workflows using LLMs, reducing manual analyst effort and accelerating insight delivery to executive stakeholders.
- ›Developed deep neural network models with autoencoder-based latent feature extraction for insurance risk prediction, improving model accuracy over traditional feature engineering approaches.
Research Assistant
Jeju National University
Jeju, South Korea
- ›Designed and implemented YOLOv3-based object detection and behavior classification models, leading to 91% performance improvement over baseline methods in video analysis.
- ›Collected, cleaned, and analyzed large web-scraped datasets to support multiple research projects.
Research Assistant
COMSATS University of Islamabad
Islamabad, Pakistan
- ›Built time series analysis and neural network models for house price prediction.
- ›Achieved 98% accuracy in energy usage prediction through hyperparameter optimization.
- ›Developed sentiment analysis pipelines using text embeddings for reviews classification.
Education
Master's in Data Science
Sapienza University of Rome
Rome, Italy | 2020 - 2024
Data Mining, Machine Learning, Statistics, Neural Networks, Cloud Computing, Deep Learning
Bachelor's in Computer Science
COMSATS University Islamabad
Attock Campus, Pakistan | 2015 - 2019
Machine Learning, Database, OOP, Data Structures, AI, Visual Programming
Publications
Face Recognition Comparative Analysis Using Different Machine Learning Approaches
Advances in Science and Technology Research Journal (ASTRJ) (2021)
Read PaperKobe Bryant Shot Prediction Using Machine Learning
Foundation University Journal of Engineering and Applied Sciences (HEC Recognized Y Category) (2021)
Read PaperPredicting Male Happiness in Pakistan: A Machine Learning Analysis of the MICS6 Dataset
Research Paper — Random Forest, SMOTE-NC, Feature Importance (2024)
Get In Touch
Have a project in mind or want to discuss an opportunity? I'd love to hear from you.
Let's work together
Whether you're looking for a data scientist to build ML pipelines, a full-stack developer to create your next web application, or a freelancer for a specific project — I'm open to opportunities.
Location
Quetta, Pakistan