Mechatronics Engineering student at Universiti Teknologi Malaysia, working at the intersection of applied AI for high-stakes systems, medical decision support, industrial diagnostics, and agentic LLM systems.
Currently:
- Aspiring Researcher
- Robotics integration on the AIROST × DF Cobot project (ROS 2, MoveIt2, Gazebo).
- Head of AI/ML at GDG on Campus UTM. NOV 2025 - FEB 2026
| Domain | Tools |
|---|---|
| Languages | Python, C/C++, ARM Assembly, SQL |
| ML / DL | PyTorch, Hugging Face, scikit-learn, XGBoost, LightGBM |
| LLM / Agents | sentence-transformers, chain-of-thought reasoning, tool-augmented agents, semantic retrieval |
| Robotics & CV | ROS 2 (Jazzy), Gazebo Harmonic, MoveIt2, ros2_control, OpenCV, Open3D |
| Tools | Git, Linux, Docker, Streamlit |
Asclepios AI — MIT Emerging Talent Program Architected the core AI engines for a clinical decision-support system for substance use disorder treatment on 1.4M patient records. Built a twin-engine segmented XGBoost framework and a Chronicity Proxy algorithm that became a top-3 predictor of relapse risk (AUC 0.75 on TEDS).
ExoDetectAI — NASA International Space Apps Challenge End-to-end ML pipeline for automated exoplanet validation on Kepler / K2 / TESS data; ensemble models (XGBoost, LightGBM, MLP) with LEO-vetter and lightkurve, deployed as a Streamlit prototype.
Neuro-Symbolic Fake News Detector Hybrid LSTM + symbolic-logic model showing improved interpretability and adversarial robustness compared to pure deep-learning baselines.
