Software Engineer and AI Researcher focused on Robust Machine Learning, Tabular AI, Computer Vision, and MLOps. My work combines empirical research with production-grade software engineering to build reliable, scalable, and reproducible AI systems. I enjoy transforming research ideas into deployable software and contributing to open-source projects.
| 📄 Elsevier Q1 Author | 🤖 AI Research Engineer |
| 🧠 Robust ML Research | 🚘 Computer Vision |
| ⚙️ MLOps & Backend | 📊 Tabular AI |
| 🌍 International Internship | 💻 Open Source |
| 👨🏫 ACM Technical Head | 🔬 Research Engineering |
| Research Principles |
|---|
| Reproducibility First |
| Engineering Before Hype |
| Evidence Over Claims |
| Production-Oriented AI |
| Open Science |
| Measurable Impact |
| Research | Engineering | Community |
|---|---|---|
| AI Research | Software Engineering | Open Source |
| ML Engineering | Backend Development | Technical Writing |
| Graduate Research | MLOps | Conference Collaboration |
| International Roles | AI Product Engineering | Mentorship |
| Software Engineering | Artificial Intelligence | Research |
|---|---|---|
| FastAPI | Deep Learning | Experimental Design |
| REST APIs | Computer Vision | Benchmarking |
| PostgreSQL | Explainable AI | Statistical Analysis |
| Docker | AutoML | Reproducibility |
| Redis | Robust Learning | Scientific Writing |
| Research Focus |
|---|
| Robust Machine Learning |
| Tabular AI |
| Computer Vision |
| Efficient Deep Learning |
| MLOps |
| Explainable AI |
| Distribution Shift |
| Synthetic Data |
| AI Infrastructure |
Robust Deep Learning for Noisy Tabular Data
Designed a dynamic loss weighting framework that improves neural network robustness against label noise, class imbalance, and distribution shifts through confidence-aware batch normalization.
| Stack | PyTorch • NumPy • OpenML |
| Scale | 20 Datasets • 10 Models |
| Evaluation | Cross Validation • Wilcoxon Tests |
| Focus | Robust Machine Learning |
| Repository | CCR-Tabular |
Highlights
- Dynamic confidence weighting
- Large-scale benchmarking
- Statistical significance testing
- Fully reproducible pipeline
Revisiting Training-Free Tabular Models
Published research benchmarking HyperFast against classical machine learning algorithms under multiple robustness scenarios, demonstrating the importance of rigorous empirical evaluation.
| Stack | Python • XGBoost • LightGBM |
| Datasets | 17 Public Benchmarks |
| Publisher | Elsevier Array (Q1) |
| Repository | CCR-Tabular |
Highlights
- Large-scale benchmark
- Robustness evaluation
- Efficient experimentation
- Publication-ready framework
Industrial Computer Vision System
Built an end-to-end Automatic Number Plate Recognition pipeline for Malaysian and ASEAN traffic environments during my industrial internship.
| Stack | PyTorch • OpenCV • YOLOv8 |
| Domain | Intelligent Transportation |
| Deployment | Industrial R&D |
| Repository | Private |
Highlights
- OCR pipeline
- Synthetic data generation
- GPU optimization
- Real-world deployment
Experiment Management Platform
A modular MLOps platform for reproducible machine learning experimentation with dataset versioning, workflow orchestration, and experiment tracking.
| Stack | FastAPI • PostgreSQL • Redis • Docker |
| Architecture | Modular Backend |
| Focus | MLOps |
| Repository | GitHub |
Highlights
- MLflow integration
- Distributed workflows
- Experiment tracking
- Dataset versioning
Explainable Malware Detection
Machine learning framework for malware classification with explainable AI techniques using SHAP feature attribution.
| Stack | Python • LightGBM • SHAP |
| Focus | Cybersecurity |
| Models | Gradient Boosting |
| Repository | GitHub |
Highlights
- Malware detection
- Explainable predictions
- SHAP visualization
- Adversarial evaluation
Cross-Platform Task Scheduler
A lightweight Python automation utility supporting scheduled task execution across Windows, Linux, and macOS.
| Stack | Python |
| Platforms | Windows • Linux • macOS |
| Focus | Automation |
| Repository | GitHub |
Highlights
- Cross-platform
- Automatic logging
- Zero configuration
- Lightweight execution
Sep 2025 - Mar 2026
- Built production-ready Computer Vision systems.
- Developed ANPR and OCR pipelines.
- Generated synthetic datasets for model training.
- Optimized deep learning models on GPU infrastructure.
PyTorch OpenCV YOLOv8 OCR Deep Learning
Apr 2026 - May 2026
- Operational analytics
- Data-driven reporting
- Workflow documentation
- Process evaluation
Research Analytics Documentation
2023
- Facial expression recognition
- Stress detection
- XGBoost models
- Automated feature engineering
Python OpenCV Machine Learning
2024 - 2026
- Organized AI workshops
- Mentored student developers
- Led technical activities
- Promoted open-source contributions
Leadership Mentoring PyTorch
| Publication | Status |
|---|---|
| Revisiting Training-Free Tabular Models Elsevier Array (Q1) |
✅ Published |
| Confidence-Calibrated Reweighting Elsevier Neurocomputing |
📝 Under Review |
- Robust Machine Learning
- Tabular Deep Learning
- Confidence-Based Loss Functions
- Distribution Shift Evaluation
- Explainable AI
- Large-Scale Benchmarking
- Reproducible ML Research
| Recognition | Details |
|---|---|
| 🏆 Elsevier Author | Published Q1 Journal Paper |
| 🔬 AI Research | Robust Machine Learning |
| 🚘 Industrial AI | Production Computer Vision |
| 🌍 International Experience | Malaysia Internship |
| 💻 Open Source | 30+ Public Repositories |
| 👨🏫 Technical Leadership | ACM Technical Head |
| 🏛 Government Research | Hyderabad City Police |
| 📄 Research Publications | Peer-reviewed AI Research |
| Role | Organization | Impact |
|---|---|---|
| Technical Head | ACM Student Chapter | AI Workshops & Mentorship |
| Open Source Maintainer | GitHub | Machine Learning Projects |
| Research Contributor | AI Community | Robust ML Research |
research:
- Robust Machine Learning
- Tabular AI
- Computer Vision
building:
- AI Infrastructure
- Open Source Tools
- MLOps Platforms
learning:
- LLM Systems
- AI Agents
- Distributed ML
open_to:
- Software Engineering
- AI Research
- ML Engineering
"Building software with engineering discipline. Advancing AI through reproducible research."