Skip to content
View MAhsaanUllah's full-sized avatar

Highlights

  • Pro

Block or report MAhsaanUllah

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
MAhsaanUllah/README.md

◼️ Muhammad Ahsaan Ullah

AI & ML Engineer | Lead @ DeepFakeShield 🛡️
Building production-ready AI systems across Computer Vision, Automation, and Data Intelligence.


◼️ Profile Summary

  • 🎓 Education: Final-year BS Computer Science (7th Semester), University of Management and Technology
  • 💻 Core Focus: Designing AI solutions for real-world security, automation, and analytics problems
  • ⚙️ Engineering Mindset: From notebooks → scalable APIs, RAG agents, and automated ETL pipelines
  • 📍 Availability: Open to Junior AI / ML / Data roles (Lahore & Remote)

◼️ Technical Pillars

  • Forensic AI
    Lead developer of DeepFakeShield, an AI system for detecting manipulated images and videos using deep learning.

  • AI Automation & RAG
    Building Retrieval-Augmented Generation agents using Gemini 2.5 Flash, vector databases, and n8n orchestration.

  • Data Analytics & BI
    Converting raw datasets into actionable insights using optimized SQL pipelines and Power BI dashboards.


◼️ Featured Repositories

🛡️ DeepFakeShield (FYP) — Lead Developer

Unified AI forensic system for identifying manipulated images and videos.
🔗 https://github.com/MAhsaanUllah

Tech Stack

  • Python, PyTorch
  • XceptionNet, EfficientNet-B0
  • FastAPI

Impact

  • Achieved 85–92% detection accuracy via frame-level analysis
  • Leveraged transfer learning for robust deepfake generalization

🤖 Enterprise RAG Support Automation

Industrial-grade RAG agent for machinery troubleshooting.
🔗 https://github.com/MAhsaanUllah/Enterprise-RAG-Support-Automation

Tech Stack

  • Google Gemini 2.5 Flash
  • Pinecone Vector Database
  • n8n Automation

Impact

  • Automates complex troubleshooting workflows
  • Reduces dependency on manual technical support

🔄 n8n Python Data ETL

Hybrid low-code + Python-based ETL pipeline.
🔗 https://github.com/MAhsaanUllah/n8n-python-data-etl

Tech Stack

  • n8n
  • Python (Pandas, NumPy)
  • JSON Processing

Impact

  • Automated data cleaning and transformation
  • Significantly reduced manual reporting effort

◼️ Tech Stack Overview

Domain Tools
Languages Python, SQL, JavaScript
AI / ML PyTorch, TensorFlow, Scikit-learn, LangChain
Computer Vision & NLP OpenCV, Hugging Face, spaCy
Automation n8n, Gemini API, Pinecone
Data & BI Power BI, SQL, Streamlit

◼️ GitHub Analytics

GitHub Stats Top Languages


◼️ Connect & Collaborate

🔗 LinkedIn
🌐 Portfolio
🐦 Twitter / X

“Build systems that work, explain them simply, and improve them continuously.”

Pinned Loading

  1. PDF-RAG-Chatbot-Gemini PDF-RAG-Chatbot-Gemini Public

    Python

  2. Retail-Sales-Forecasting-Pipeline Retail-Sales-Forecasting-Pipeline Public

    Jupyter Notebook

  3. ai-tool-calling-agent ai-tool-calling-agent Public

    An autonomous AI Agent built with n8n and Google Gemini that uses Multi-Step Reasoning to execute external tools (Calculator, Wikipedia) for solving complex queries.

  4. n8n-python-data-etl n8n-python-data-etl Public

    A robust ETL pipeline using n8n and Python to sanitize raw JSON data, demonstrating hybrid low-code orchestration with custom script-based data transformation.

    Python

  5. desktop-media-ai-organizer desktop-media-ai-organizer Public

    Automatically organize your desktop media files (images, memes, screenshots, documents) using AI-powered classification and a user-friendly interface.

    Python

  6. GestureMouseCV GestureMouseCV Public

    Control your PC with hand gestures via webcam — real-time Computer Vision pipeline using MediaPipe, OpenCV & TensorFlow. No mouse required.

    Python