BS Data Science & AI — FAST NUCES, Lahore
Building end-to-end machine learning systems, data pipelines, and backend-driven applications.
Data Science undergraduate with hands-on experience building deployable AI systems across fraud detection, computer vision, NLP, and data engineering. I focus on evaluation rigor, clean architecture, and solutions that work in real scenarios — not just in notebooks.
| Domain | Technologies |
|---|---|
| Languages | Python, C++, C#, SQL |
| ML / AI | Scikit-learn, PyTorch, XGBoost, NLP, LLMs, RAG, PEFT, SHAP, Prompt Engineering |
| Data Engineering | Pandas, PostgreSQL, SQL Server, ETL Pipelines, Data Warehousing, Power BI |
| Backend & Tools | Streamlit, .NET, Jupyter, VS Code, Git |
| Cloud | AWS, Microsoft Azure |
- Built ML pipelines for fraud detection, recommendation, and forecasting with measurable evaluation metrics
- Developed backend and database-driven systems handling authentication, workflows, and structured data
- Designed ETL pipelines, data warehouse schemas, and Power BI dashboards for business reporting
- Integrated LLM-based features including RAG systems, PEFT fine-tuning, and prompt engineering
Problem understanding → data preparation → model or system design → implementation → evaluation → delivery
I focus on solutions that are not only technically correct but also reproducible, maintainable, and useful in real scenarios.
- Applied AI & ML Products: Fraud detection, recommendation, and forecasting systems built with rigorous evaluation (AUPRC, recall, SHAP explainability) and deployed via Streamlit or HuggingFace.
- Data Engineering & Analytics: ETL workflows, star-schema warehouse designs, and reporting layers that turn raw transactional data into actionable dashboards.
- Backend & Database Systems: Structured backend logic, authentication flows, admin approval pipelines, and SQL-backed application features.
- End-to-End Prototyping: From concept to deployable demo with clear documentation, modular code, and maintainable architecture.
| Project | What I Built | Stack | Highlights |
|---|---|---|---|
| TrustGuard AI | Fraud detection pipeline on PaySim with explainable outputs and regulatory grounding | Python, XGBoost, PyTorch, RAG, ChromaDB | 4 models, SMOTE for 0.13% class imbalance, SHAP explainability, RAG grounded in SBP regulations, deployed on Streamlit |
| PEFT Comparative Study | Parameter-efficient fine-tuning analysis across multiple LLM adaptation methods | Python, PyTorch, HuggingFace, PEFT | LoRA-based tuning pipelines, multi-seed statistical evaluation, McNemar significance testing |
| Time Series Data Analysis & Trend Discovery in Pakistan Crop Prices | End-to-end time-series forecasting and anomaly detection on 53 CSVs | Python, Pandas, Scikit-learn | 9 models benchmarked; Linear Regression ranked first by RMSE; lag_1 identified as dominant universal predictor |
| Harris-LK Object Tracker | Classical single-object tracker combining corner detection and optical flow | Python, OpenCV, NumPy | Harris + pyramidal LK, forward-backward error filtering, adaptive redetection; 54-page technical report |
| Movie Recommendation System | Hybrid recommendation app with interactive interface | Python, Streamlit, Scikit-learn | Content-based + collaborative filtering via cosine similarity |
| E-Commerce Data Warehouse & Analytics | Reporting and analytics solution for business insights | PostgreSQL, Power BI | Star schema design, ETL from raw transactional data, Power BI dashboards |
| Mock Examination System | Digital mock-test platform with structured exam workflows | Python, SQL | Timed attempts, scoring engine, result reporting, SQL-backed schema |
| Unused Medicine Donation System | Full-stack platform for medicine donation and request workflows | C#, .NET, SQL | Auth, admin approval pipeline, donor-recipient matching |
Specializations
| Certificate | Issuer | Date |
|---|---|---|
| Prompt Engineering Specialization | Vanderbilt University | Aug 2025 |
| Generative AI Assistants Specialization | Vanderbilt University | Aug 2025 |
| Google AI Essentials | Jul 2025 | |
| Google Prompting Essentials | Jul 2025 | |
| Generative AI for Educators | IBM | Jul 2025 |
Individual Courses (5)
| Certificate | Issuer | Date |
|---|---|---|
| Working with the OpenAI API | DataCamp | Nov 2025 |
| Prompt Engineering with the OpenAI API | DataCamp | Nov 2025 |
| Feature Engineering for ML in Python | DataCamp | Oct 2025 |
| Introduction to Deep Learning with PyTorch | DataCamp | Oct 2025 |
| AI For Everyone | DeepLearning.AI | Jul 2025 |
GitHub • LinkedIn • X • Instagram • Linktree
"The gap between an idea and a working system is where I live."