Resume AI/ML Engineer in Pakistan Faisalabad

Research Associate Computational Drug Design Lab Focus: Pediatric Neurological Comorbidities Sep '23 — Present Islamabad, Pakistan Investigated the comorbidity between ADHD and ASD to identify overlapping and disease-specific approved drug targets. Conducted extensive data wrangling and manipulation on large-scale complex datasets related to neurodevelopmental disorders Implemented Python scripts to automate the batch Molecular Docking and protein-ligand interaction analysis using PLIF. Built and trained Machine Learning and Deep Learning models to predict drug-target affinites. Focused on accelerating and enhancing the drug screening process against approved therapeutic targets.

Master's in Bioinformatics, National University of Sciences & Technology, NUST Islamabad

Core Skills: Data Science, Deep Learning, Machine Learning, Scripting, Computational Drug Design Programming Languages: Python, R (Scikit-learn, TensorFlow, Keras, Biopython, RDKit, XGBoost, Pandas, NumPy, Matplotlib)

English

PROJECTS: 1. Waste Classification Through Convolutional Neural Network - Collected a labeled waste image dataset from Kaggle to train and validate the model. - Designed a CNN for real-time classification with high accuracy and low computational overhead. - Compared model performance against benchmark architectures (ResNet50 and MobileNet) to validate efficiency and precision. - Built an interactive Streamlit web application to deploy the model, enabling users to upload waste images and receive instant classification feedback. - Achieved 83% accuracy in distinguishing between organic and recyclable waste, promoting sustainable waste management. - The model can be integrated into smart waste bin systems to assist in proper waste disposal and encourage recycling behavior. PROJECT 2. Revolutionizing Healthcare with Early Heart Disease Mortality Prediction - Designed a machine learning-based system to predict early mortality risk in heart disease patients, aiming to support timely clinical interventions and improve patient outcomes. - The project utilized a real-world, publicly available dataset from the Institute of Cardiology, Faisalabad, which included key patient attributes and clinical indicators. - Developed and evaluated multiple machine learning models, including Logistic Regression, Decision Tree, and Random Forest. - Random Forest outperformed other models, achieving the highest prediction accuracy of 96%. - Provided insights into key risk factors contributing to early mortality, aiding in data-driven decision-making for healthcare professionals. PUBLICATION: Biosynthesis of Fosfomycin Loaded CuO Nanoparticles: Evaluation of Antibacterial, Antibiofilm properties and Molecular Docking Analysis against Biofilm Associated Proteins in MDR bacteria Journal of Organometallic Chemistry (Elsevier) Impact factor 2.1 Under Review Interests: Medical Image Analysis