Skip to content
View tgishor's full-sized avatar
🎯
Focusing
🎯
Focusing

Highlights

  • Pro

Organizations

@Macquarie-University-MIT-AI-2025

Block or report tgishor

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
tgishor/README.md

Hi, I'm Gishor Thavakumar

🎯 AI Engineer • ML Researcher • Data Scientist

Building intelligent systems that solve real-world problems
🎓 Master's in AI • 💼 2+ Years Industry Experience • 🌏 Open to Opportunities

Website LinkedIn HuggingFace


🚀 About Me

Recent Master of IT (Artificial Intelligence) graduate from Macquarie University with hands-on experience as a Business Intelligence Analyst at Techants Australia. I build AI systems that work in production.

Currently fascinated by embedded AI - making intelligent systems work on resource-constrained hardware.


💼 Experience

Business Intelligence Analyst @ Techants Australia Oct 2023 - Aug 2025

  • Developed end-to-end ML pipelines processing millions of data points
  • Built predictive models that improved decision-making across business units
  • Created automated dashboards reducing manual reporting by 80%

Master of IT (AI) @ Macquarie University Jul 2023 - Aug 2025

  • Research focus: Deep Learning, Computer Vision, Graph Neural Networks

Languages & Frameworks

AI/ML & Data Science

Robotics & Embedded Systems

Databases & Cloud

Development Tools

Statistical & Business Intelligence


🎯 Featured Projects

Embedded AI • Computer Vision • Real-time Processing

Real-time movement coordination system for robotic choreography using edge computing and computer vision.

BERT • RoBERTa • PyTorch • Production-Ready

High-performance text classification using state-of-the-art transformer models with optimized inference.

ResNet50 • CBAM • Fine-grained Classification

Precision aircraft identification using attention-enhanced CNN architectures achieving SOTA results.

GraphSAGE • GAT • Social Networks

Friend recommendation system using graph neural networks with 23% accuracy improvement over baselines.

🔍 View All Projects

AI & Machine Learning

Analytics & Business Intelligence

Data Science & Statistical Analysis

Full-Stack Development


📈 GitHub Analytics


💬 Let's Connect

Ready to discuss how AI can solve your toughest challenges

Portfolio LinkedIn


Pinned Loading

  1. Multi-Robots-Choreography-Controller Multi-Robots-Choreography-Controller Public

    Advanced ROS2 robotics platform that transforms music into synchronized dance performances using real-time audio analysis, algorithmic movement generation, and multi-robot coordination. Features Me…

    Python 1

  2. Macquarie-University-MIT-AI-2025/Catinator Macquarie-University-MIT-AI-2025/Catinator Public

    A fine-grain classifier for cat breeds, 21 breeds were explored and a robot was programmed to perform actions based on a successful prediction.

    Python 1 2

  3. FGVC-Aircraft-Manufacturer-and-Variant-Classification-with-ResNet50-and-CBAM-PyTorch FGVC-Aircraft-Manufacturer-and-Variant-Classification-with-ResNet50-and-CBAM-PyTorch Public

    Fine-grained aircraft image classification on the FGVC-Aircraft dataset at manufacturer and variant levels using PyTorch. Compares ResNet-50 (transfer learning & fine-tuning) with a custom CBAM att…

    Python

  4. Conditional-VAE-for-MNIST-Digit-Reconstruction-and-Representation-Learning-PyTorch Conditional-VAE-for-MNIST-Digit-Reconstruction-and-Representation-Learning-PyTorch Public

    Conditional VAE in PyTorch for MNIST digit reconstruction & controlled generation. Includes ELBO loss, importance sampling, data augmentation, t-SNE latent space visualization, and performance eval…

    Jupyter Notebook

  5. Friend-Recommendation-Using-GNNs-GraphSAGE-GAT-PyTorch Friend-Recommendation-Using-GNNs-GraphSAGE-GAT-PyTorch Public

    Compares GraphSAGE and GAT in PyTorch (PyG) to suggest top-K friends on a social graph. Includes data prep & edge splits, mini-batch training with negative sampling, evaluation (ROC-AUC, AP, Hits@K…

    Jupyter Notebook

  6. Medical-QA-LSTM-Transformer-Models-Comparison Medical-QA-LSTM-Transformer-Models-Comparison Public

    AI-powered medical question answering system comparing Siamese, LSTM & BERT neural networks on BioASQ dataset. Achieved 83% performance improvement with LSTM approach (F1: 0.42). Demonstrates deep …

    Jupyter Notebook