Languages
Python · primary
SQL
JavaScript · basic
HTML
CSS
Dart
AI INTERN · BSCS · SEMESTER 7
Mahad Khan
AI Intern
I build systems that turn raw data into decisions — from graph-based anomaly detection to the retrieval and evaluation logic behind modern AI applications.
- Focus
- AI & Data Systems
- Core Tool
- Python
- Learning
- PyTorch
01 · About Me
Curious about how data becomes intelligence.
I'm a 7th semester BSCS student at COMSATS University Islamabad, Wah Campus, majoring in Artificial Intelligence and Data Science. I like taking messy, real-world data and turning it into models and pipelines that actually explain something — whether that's tracing fraud through a transaction graph or designing the schema that lets an application retrieve the right context fast.
My current focus sits at the intersection of Machine Learning, database design, and Generative AI — I'm exploring retrieval-augmented generation, embeddings, and evaluation techniques like hallucination detection, alongside the core CS foundations (data structures, algorithms, DBMS) that make those systems reliable. I'm actively looking for opportunities to apply this as an AI / Data Science practitioner.
- EducationBSCS, COMSATS University Islamabad — Wah Campus
- Standing7th Semester
- InterestsMachine Learning, Graph Algorithms, RAG & LLM Evaluation, Databases
02 · Skills
What I work with.
AI / Generative AI
LLMs
Prompt Engineering
RAG Concepts
Few-Shot Learning
Embeddings
Vector Search
Hallucination Detection · conceptual
OpenAI APIs
ML / Data
Machine Learning
Linear Regression
Graph Algorithms (DFS)
NumPy
Pandas
Matplotlib
NetworkX
PyTorch · learning
Databases
MySQL
MongoDB
PostgreSQL
Supabase
Relational & NoSQL Design
pgvector · conceptual
Vector DB Concepts
Core CS
Data Structures & Algorithms
Object-Oriented Programming
DBMS
Software Engineering
Cybersecurity Fundamentals
03 · Projects
Selected academic work.
Graph-Based Machine Learning
2025
Financial Fraud Pattern Mining
- Designed a graph-based anomaly detection system to identify fraudulent transaction patterns using Depth-First Search traversal on transaction networks — aligned with multi-step reasoning and LLM evaluation pipeline concepts.
- Applied Linear Regression for trend analysis and used NetworkX to model transaction graphs, visualizing findings with Matplotlib to demonstrate an end-to-end data pipeline.
- Explored anomaly detection and output-quality measurement logic applicable to hallucination detection and LLM evaluation in Generative AI systems.
Python
NumPy
Pandas
NetworkX
Matplotlib
Advanced Database Project
2025
Hotel Management System
- Built a full-stack, web-based hotel management system with modules for room booking, guest records, billing, and inventory management.
- Designed MongoDB collections and schema, managing relationships through NoSQL modeling — foundational skills for scalable AI service backends and context-handling data layers.
- Structured data retrieval logic for efficient querying, laying groundwork for semantic search and vector-based retrieval with tools like pgvector and Azure AI Search.
MongoDB
HTML
CSS
JavaScript
04 · Contact
Let's build something with data.
Open to internships, collaborations, and conversations about AI, machine learning, and data engineering. Reach out directly or send a message below.