Open to opportunities

Abdullah User

@abdullahuser4

Message

Junior Machine Learning Engineer skilled in Python and deep learning.

Zimbabwe

Message

What I'm looking for

I am looking for opportunities to grow in a collaborative environment that values innovation and continuous learning.

I am a Junior Machine Learning Engineer with a strong foundation in Python and deep learning techniques, including natural language processing (NLP) and computer vision. My experience encompasses building, training, and deploying machine learning models, with a focus on data preprocessing, model optimization, and evaluation. I am particularly knowledgeable in transformer architectures and attention mechanisms, which I have applied in various projects.

Throughout my academic journey, I have successfully implemented advanced projects such as a Reformer Transformer from scratch, achieving significant memory efficiency improvements. Additionally, I developed a financial sentiment analysis model using FinBERT, which attained an impressive accuracy of 98.9%. My commitment to continuous learning and collaborative development in AI projects drives my passion for creating innovative solutions that address real-world challenges.

Experience

Work history, roles, and key accomplishments

Financial Sentiment Analysis

Self Employed

Built a real-time NLP model using FinBERT, achieving 98.9% accuracy on financial news sentiment. Created a complete ML pipeline including preprocessing, tokenization, fine-tuning, evaluation, and a REST API with Flask and Docker.

NLP FinBERT Flask Docker REST APIs Machine Learning Pipelines

Reformer Transformer From Scratch

Self Employed

Implemented Reformer architecture in PyTorch with LSH attention and reversible layers for memory-efficient training on long sequences. Demonstrated advanced transformer optimization, reducing memory from quadratic to sub-quadratic complexity.

PyTorch Deep Learning Neural Networks

Intrusion Detection System

Self Employed

Developed an Intrusion Detection System using Random Forest and XGBoost with 95.9% accuracy. Integrated SHAP for explainability and feature importance visualization, packaging the system with FastAPI endpoints and Docker for deployment.

Random Forest XGBoost SHAP fastAPI Docker Intrusion Detection Machine Learning Explainable AI

GPT-2 Fine-Tuning

Self Employed

Fine-tuned GPT-2 (117M) on a sci-fi corpus to generate coherent futuristic narratives. Employed Hugging Face Transformers with PyTorch and GPU acceleration on Colab, tuning hyperparameters with AdamW optimizer.

PyTorch GPU Acceleration Hyperparameter Tuning NLP Text Generation

Education

Degrees, certifications, and relevant coursework

Sylhet Science and Technology College

Higher Secondary Certificate (HSC), Science Stream

2024 -

Pursuing a Higher Secondary Certificate in the Science Stream. This program focuses on foundational scientific principles and prepares students for advanced studies.