"Real-world AI isn't about the highest accuracy on a clean dataset — it's about data quality, reliable pipelines, and products that actually work outside a notebook."
About · Tech Stack · Projects · Experience · Stats
I'm a 4th-year Information Technology student at the Faculty of Computers and Artificial Intelligence (graduating 2027), building my path across Artificial Intelligence, Data Engineering, and Machine Learning.
My journey started with Business Intelligence — using SQL, Python, Power BI, and Tableau to turn messy raw data into clear insights. I quickly realized I wanted to build systems, not just dashboards, which pushed me deeper into machine learning and data engineering.
- 🧭 Currently a Data Engineer Intern @ DEPI (Microsoft Track) — working with Azure, big data processing, and scalable data pipelines.
- 🎓 Recently completed ITI's Machine Learning Level 1 track. For the final project, I built an end-to-end real estate pricing product on 180,000+ listings.
- 🔬 I enjoy tackling non-traditional data: face detection with neural networks, rock/mineral image classification, and emotional speech recognition.
- 🚀 Currently taking ITI's Machine Learning Level 2 (LLMs, RAG, Computer Vision) and leading Team Nexora at the FortyGuard Hackathon '26.
I'm always learning, always building, and always looking for the next problem worth solving.
🤖 AI & Deep Learning
📊 Data Engineering & BI
🌐 Backend & Web
🪐 Orbit Wars — Kaggle RL Competition (Rank 102)
An optimized, PyTorch-based RL agent built for the Kaggle Orbit Wars strategy competition, leading team NEXUS to a final rank of 102 with a top score of 1252. Uses a vectorized trajectory/intercept engine, a custom "Enemy Pressure" threat model, and continuous collision avoidance.
Python PyTorch RL ➔ View Repository
🌡️ CoolPriority — FortyGuard Hackathon '26
AI-powered urban heat decision-support system. Combines FortyGuard hyperlocal temperature data and the CDC/ATSDR Social Vulnerability Index into a normalized Cooling Priority Score through an interactive Streamlit dashboard.
Python Streamlit Geospatial Analysis ➔ View Repository
🏠 House Price Prediction — End-to-End ML App
An ML-powered real estate pricing product built on 180,000+ real property listings. XGBoost was selected for its balance of predictive performance and deployment practicality, served through a FastAPI backend with a React + TypeScript frontend.
XGBoost FastAPI React TypeScript ➔ View Repository
🗣️ Emotional Speech Recognition (99.29% Accuracy)
Classifies 2,800 audio files into 7 emotions using a custom 60-dimensional audio feature extraction pipeline. Compares MLP and SVM classifiers.
Python Scikit-Learn Audio Features ➔ View Repository
✋ Face Detection — YOLOv8n Pipeline
A highly optimized face detection pipeline built on the anchor-free YOLOv8n architecture, tuned to balance speed and accuracy in real-time scenarios.
Python YOLOv8 OpenCV ➔ View Repository
- 🔷 Data Engineer Intern — DEPI (Microsoft Track)
(Jul 2026 — Present)- Building practical foundations in data engineering, data pipelines, big data processing, and Microsoft Azure.
- 🔷 Machine Learning Trainee — ITI
(Jul 2026 — Aug 2026)- Completed ML Level 1. Built an end-to-end house price prediction application on 180,000+ listings from scratch to deployment.
- 🔷 Business Intelligence Trainee — ITI
(Sep 2025 — Present)- Applied Data Warehousing concepts; built interactive dashboards (Power BI/Tableau) and advanced SQL/SSIS pipelines.
- 🎓 B.Sc. Information Technology
(Expected 2027)- Faculty of Computers and Artificial Intelligence.
