Master Degree in AI for Science

This is a unique Master's program at AIMS South Africa in collaboration with Google DeepMind. It offered me a rigorous curriculum that explored the dynamic intersection of AI and the Sciences. I have engaged in a diverse range of courses, including Introduction to Machine Learning, Applied ML at Scale, Formal Methods and Functional Programming, Computer Vision, Statistical Inference and Causality, CUDA/GPU, Simulation-based Inference, ....
My thesis project, supervised by the Centre for Epidemic Response and Innovation (CERI), focused on 'AI-Enhanced Classification and Early Detection of SARS-CoV-2 Variants for Timely Public Health Response'. This research carried under the supervision of Dr. Joicymara Xavier and Dr. Houriiyah Tegally represents a proactive approach to potentially mitigating the impact of future pandemics, ensuring society is better prepared for future health crises.

Engineering Degree in Mathematical and Modeling Engineering

Through this training, I delved into applied mathematics, statistics, optimization and programming. I also acquired some practical experience in data analysis and machine learning through my thesis project focused on predicting the coastline evolution based on satellite images.

Preparatory classes for engineering school

Through this rigorous two-year preparatory program, I completed a comprehensive curriculum designed to prepare students for advanced studies in engineering. The program included specialized coursework in Mathematics for Engineers, covering a range of topics from foundational principles to advanced applications (analysis, algebra, ODEs, PDEs, Laplace & Fourier transforms , etc). Additionally, I gained proficiency in algorithmics, graphs, optimization and programming providing a solid foundation for subsequent engineering training.

Bioacoustic recognition technology for Zoonotic disease preparedness

AI-enhanced classification and early detection of SARS-CoV-2 variants for timely public health response

• Implemented state-of-the-art genomic feature extraction techniques including k-mers and FCGRs.
• Designed and evaluated both flat and hierarchical at categorizing SARS-CoV-2 lineages based on extracted genomic features.
• Proposed an active learning approach to improve the detection of emerging variants.
• Achieved timely detection of new variants, consistently identifying them within 15 days of their first appearance.

Coastline position prediction from satellite images: comparison of statistical and machine learning models

• Extracted coastline position time series data from satellite images;
• Cleaned the data;
• Built a LSTM model to predict the coastline position;
• Built a ARIMA model to predict the coastline position;
• Compared the performance of the two models.

Maize disease classification using CNNs and plant leaf images.

• Cleaned the data;
• Built a CNN model to predict the disease type;
• Fine-tuned the VGG16 model to predict the disease type;
• Compared the performance of the two models.

Skin phototype prediction using facial images, ANNs & SVM.

• Calculated texture features on skin pieces;
• Trained a model to cluster the input images in the features space;
• Extracted piece of skin from facial images;
• Trained an ANN and a SVM models to predict the phototype.

Data-driven Life Sciences course 2024

• Explored the intersection of data science, AI, and life sciences.
• Covered application areas such as proteomics, transcriptomics, biomolecular structure, molecular dynamics simulations, and imaging techniques.
• Full list of attended lectures
Introduction to Data-Driven Life Science and AI, Protein structure prediction and design, Large-scale predictions of antibiotic resistance in microbial communities, Decoding Cell-Level Biology with scRNA-Seq Data, Metagenomics and Microbiome Analysis, Advances in Generative AI for Conditional Microscopy Cell Images, Deep learning enabled image-driven scientific discovery, From Single Molecules to Tissue Atlases, System Biology and computational modeling, Accelerating Scientific Discovery with Large Language Models.

2024 EAUMP-ICTP School on the Mathematics of AI

• Explored foundational mathematics of AI, including data assimilation, inverse problems, and machine learning;
• Engaged with experts in diverse fields such as control theory, graph theory, and statistics;
• Full list of attended lectures :
Sequential Decision Making, Graph tools for AI and ML, R and machine learning, Statistics of AI, Probabilistic machine learning, Data analytics and Machine Learning and Foundations of Neural Networks and Deep Learning.