⚡ Machine Learning & Parallel Computing (MLPC)

This repository contains both major assessments completed for the Machine Learning & Parallel Computing (ITS66604) module.

Across these projects, I explored the foundations of supercomputing, GPU architecture, parallel processing, and machine learning workflows, applying them to real-world analytical tasks.

The repository includes:

• 🖥 High-Performance Computing (HPC) Supercomputer Analysis
• 🤖 Machine Learning Classification Project (Weather Extreme Detection)

---

🚀 What I Learned in This Module

🧠 Machine Learning Foundations

• Exploratory Data Analysis (EDA)
• Data preprocessing & cleaning
• Handling anomalies and missing values
• Label engineering (custom rule-based classification)
• Feature engineering (interaction features, distractor features)
• Gaussian noise injection for robustness
• Class imbalance handling (oversampling)
• Encoding & scaling techniques (LabelEncoder, StandardScaler)
• Machine learning workflows using:
  • Decision Tree
  • Random Forest
• Model evaluation using:
  • Confusion matrix
  • Precision/Recall/F1-score
  • ROC Curve & AUC scores
    (Random Forest AUC = 0.99 — excellent performance)

---

🔥 Parallel Computing & HPC Concepts

From the Supercomputer project:

• Supercomputer architecture (NVIDIA DGX A100 + SuperPOD)
• Parallel GPU processing (CUDA cores, Tensor Cores)
• NVLink + NVSwitch high-speed interconnect
• Distributed computing & InfiniBand networking
• GPU virtualization (MIG)
• HPC benchmark interpretation (HPL, HPCG, MLPerf)
• Storage hierarchy & parallel file systems
• Energy-efficient computing (Green500 insights)
• Use cases:
  • AI training at scale
  • Autonomous driving simulation
  • Weather & climate modeling
  • Scientific computing & medical research

---

📘 Projects in This Repository

🖥 1. Supercomputer Architecture Analysis (NVIDIA DGX A100 + SuperPOD)

A complete study of HPC design, including:

• GPU/CPU topology
• NVLink / NVSwitch connectivity
• InfiniBand networking
• Multi-node scaling
• Power & cooling strategies
• AI-accelerated workloads

📁 Location: /Supercomputer/

---

🌦 2. Extreme Weather Classification (Machine Learning)

A full ML pipeline to classify Normal vs Extreme weather using global meteorological data.

Includes:

• 40+ weather features
• Full preprocessing pipeline
• Noise injection & advanced feature engineering
• Oversampling to fix class imbalance
• Decision Tree vs Random Forest comparison
• Achieved up to 95% accuracy and AUC 0.99

📁 Location: /Extreme Weather Classification/

---

🧩 Skills Demonstrated

🔧 Technical Skills

• Machine Learning (scikit-learn)
• Data cleaning & feature engineering
• EDA visualization (Seaborn, Matplotlib)
• GPU architecture & parallel computing
• HPC system design & benchmarking
• Jupyter/Colab development
• CSV data processing at scale

📊 Analytical Skills

• Performance comparison & interpretation
• Model explainability (feature importance)
• Handling real-world ML challenges (noise, imbalance, anomalies)

📝 Professional Skills

• Technical reporting
• Academic poster design
• Structured documentation
• Scientific analysis & evaluation

---

✨ Thank you for exploring my MLPC repository!