Palm Tree Detection Using YOLOv8

● Developed an object detection model using YOLOv8 to identify and localize palm trees in aerial and drone imagery. Trained on custom-annotated datasets with data augmentation to improve accuracy. The solution supports automated plantation mapping and palm count estimation. Tools: Python, YOLOv8, PyTorch, OpenCV, LabelImg.

McDonald's Outlet Locator with Chatbot

● Built a full-stack web application that maps McDonald’s outlets in Malaysia with 5KM catchment areas using geospatial data. Implemented a state-based filtering system and a chatbot powered by local Retrieval-Augmented Generation (RAG) to answer outlet-related queries (e.g., 24-hour services, birthday party availability).

Law Employment Chatbot with RAG and Local LLM Inference

● This chatbot helps users understand Malaysian employment laws (Employment Act 1955 & Amendment 2022) by answering questions using a Retrieval-Augmented Generation (RAG) system and a locally hosted LLM. Key Highlights: ● Document Embedding: Legal texts are chunked and embedded using SentenceTransformers. ● FAISS Index: Enables fast semantic search of relevant law sections. ● LLM Response: Uses Nous-Hermes-2-Mistral-7B-DPO via llama-cpp-python for local, private inference. ● FastAPI Backend + Streamlit Frontend: Clean, interactive UI with API integration. ● Offline-Ready & Secure: No external calls; ideal for privacy-sensitive legal environments.