Yong Kheng Beh

Upcoming role in contributing to the IOC-UNESCO Ocean Decade Programme as a Data Scientist, focusing on data cleaning, text mining, and data visualisation tasks related to global sustainability initiatives.

Key Skills: Python, Data Cleaning, Text Mining, Data Visualisation, Remote Collaboration

Key Achievements:

• Selected for a volunteer position in a prestigious international programme, demonstrating strong data science skills and commitment to global sustainability.
• Anticipated to contribute to the development of data-driven insights for ocean sustainability projects.

Contributing to the rebuild of SAcommunity.org, a not-for-profit organisation website, as part of a cross-functional team using Drupal 11.

Key Skills: Drupal 11, PHP, MySQL, Agile Software Development, Team Collaboration

Key Achievements:

• Authored internal documentation and onboarding guides to support knowledge sharing.
• Self-taught new system architecture to contribute effectively to web migration tasks.

Developed automated evaluation software to compare the semantic search performance of different embedding models, chunking methods, and vector stores. Conducted extensive research and provided detailed technical documentation to facilitate future experimentation.

Key Skills: Python, Bash (Linux), AI/ML (Embeddings, Vector Search), Docker, OpenSearch

Key Achievements:

• Developed a Python-based testing framework, automated via shell scripts to streamline semantic search evaluation, significantly reduced manual testing time.
• Conducted extensive experimentation and benchmarking with top-ranked embedding models from the MTEB leaderboard on Hugging Face.
• Benchmarked and optimized semantic search performance across multiple vector stores, including OpenSearch.
• Resolved a critical semantic ranking bug, identifying that the system used cosine distance instead of similarity scoring, which significantly improved search accuracy and result relevance.
• Documented experimentation results and written comprehensive technical reports to ensure effective knowledge transfer.

Led backend development for a full-stack AI-powered chatbot as part of a four-member team for SAcommunity.org, a community directory website.

Key Skills: Python, LangChain, Flask, Docker, Agile Software Development

Key Achievements:

• Led backend development, successfully integrating the LangChain library to enable LLM-based tool calls and semantic search, resulting in an Agentic RAG (Retrieval-Augmented Generation) chatbot capable of handling vague human queries with high accuracy.
• Optimized chatbot performance by applying advanced prompt engineering and retrieval strategies to improve response relevance.
• Deployed the chatbot using Docker, resolving compatibility issues within an outdated software environment and ensuring smooth operation.
• Worked closely with frontend developers and stakeholders to refine chatbot functionality for real-world use cases.

Awarded AIML Summer Research Project Scholarship, applying a unique blend of medical and computing knowledge to advance AI applications in cardiology

Key Skills: Python, Data Analysis(Pandas, Matplotlib, Seaborn), Machine Learning(Scikit-Learn)

Key Achievements:

• Collaborated with healthcare experts and machine learning specialist to explore AI-driven decision-making solutions for medical applications.
• Processed and analyzed a large-scale healthcare dataset containing 6,000+ patient records, developing machine learning models to predict patient outcomes.
• Researched and reviewed AI-driven predictive analytics in healthcare, integrating insights into model design.
• Presented findings at the end of the internship to a panel of researchers.