Roy Ho

Experience

1. 

   JAIKE

   Journal of Artificial Intelligence and Knowledge Engineering

   January 2025 – Present

   Artificial Intelligence Researcher

   • Conducted research on LLM-based automation covering API-driven system design, agent orchestration frameworks, retrieval-augmented generation (RAG), and a consolidated threat model spanning prompt injection, data leakage, and unintended action execution.
   • First-authored a 30-page, 36-source review with Dr. Biren (Brian) Prasad as co-author, introducing a two-axis framework for LLM agent autonomy and oversight, plus a cost model pricing oversight against productivity gains; revise-and-resubmit with publication recommended.
   • Served as a peer reviewer for JAIKE, evaluating submissions on retrieval methods in large language models, reasoning performance in extended tasks, and large-scale model architectures for methodological rigor and evaluation quality.
2. 

   TechSprint Innovators

   March 2024 – September 2025

   Head of Data Engineering

   • Built a multi-factor stock screening model using fundamental, technical, and NLP-based sentiment features (FinBERT).
   • Developed and evaluated a supervised classification model in scikit-learn to predict price appreciation, performing feature selection, model tuning, and performance validation on historical market data.
   • Engineered and automated a daily ETL data pipeline (Python, yfinance, Alpaca API) running on a Raspberry Pi to filter equities, generate structured CSV outputs, and deliver real-time investment signals via Discord webhook.
3. 

   AISC

   AI Student Collective

   September 2024 – April 2025

   General Member

   • Built machine learning models for stroke risk prediction and real-time drowsy driver detection within structured project cycles.
   • Performed data preprocessing, feature engineering, model development, evaluation, and project presentations.
   • Worked within quarter-long sprint cycles with defined milestones, code reviews, and final project demos, following structured machine learning development workflows from ideation to deployment.