Hi, I'm Khalid Sabban

Passionate about LLMs and AI's power to transform lives. Building impactful ML systems that make a difference.

10+
Projects Completed
5+
Research Papers
50+
Kaggle Contributions
1k+
GitHub Stars

About Me

Building the future with AI, one model at a time

I'm a Junior Data Scientist and Machine Learning Engineer specializing in Large Language Models and cutting-edge AI applications.

My journey in AI is driven by a deep belief in its transformative power. From fine-tuning transformers to deploying production-ready ML systems, I focus on creating solutions that make a real impact.

Currently exploring transformer quantization and edge deployment of LLMs, pushing the boundaries of what's possible with AI.

๐Ÿค–

Technologies & Tools

My technical arsenal for building intelligent systems

๐Ÿ
Python
๐Ÿ”ฅ
PyTorch
๐Ÿ“Š
TensorFlow
๐Ÿค—
HuggingFace
๐Ÿณ
Docker
โšก
FastAPI
๐Ÿฆœ
LangChain
๐Ÿง 
Transformers

Featured Projects

Building innovative AI solutions that push boundaries

๐ŸŽง

Spotify Popularity Predictor

End-to-end ML pipeline to predict track popularity. Features data collection (Spotify API), NLP on playlist names (NLTK), and feature importance analysis (XGBoost).

XGBoost Scikit-learn NLP API
๐ŸŽต

Drake Lyric Generator

Character-level language model using GPT-2 architecture to generate Drake-style lyrics. Learns patterns, style, and structure through deep learning.

PyTorch Transformers NLP
๐Ÿ’พ

Llama-2 SQL Generator

Fine-tuned Llama-2 7B model for natural language to SQL query generation. Achieved 0.19 training loss after 5 epochs on specialized dataset.

Llama-2 Fine-tuning SQL
๐Ÿ”

Attention Is All You Need

Complete PyTorch implementation of the groundbreaking transformer paper from scratch. Deep dive into attention mechanisms and architecture.

PyTorch Transformers Research
๐Ÿ’ป

Laptop Price Predictor

ML model with interactive dashboard for predicting laptop prices based on specifications. Features comprehensive EDA and model comparison.

Scikit-learn Streamlit Regression
๐Ÿšข

Titanic Survival Analysis

Classic Kaggle competition solution with advanced feature engineering and ensemble methods. Comprehensive analysis of survival patterns.

Kaggle Classification EDA
โœˆ๏ธ

Airbnb Destination Predictor

ML pipeline predicting first booking destinations for new users based on behavioral patterns and demographics.

Random Forest Feature Engineering Multi-class

Let's Build Something Amazing

Open to collaborations, research opportunities, and innovative projects