Independent ai researcher - Self Employed
(2023-04)
Conducted independent research and developed machine learning solutions in quantitative finance, computer vision, NLP, speech recognition, reinforcement learning, and geospatial AI, achieving notable results in competitive research challenges and self-directed projects
- World Bank Poverty Prediction Challenge: Ranked 13th of 1,275 teams (Top 1%), developing robust ensemble models emphasizing feature stability and generalization
- TAHMO Incoming Solar Radiation Challenge: Ranked 74th of 1,529 participants (Top 5%), earning a Zindi Silver Medal for forecasting nonlinear sensor drift in meteorological time-series data
- Hull Tactical Market Prediction: Developed a reinforcement learning-based trading agent for risk-aware portfolio optimization, achieving a Top 25 finish with a Sharpe Ratio of 1.486
- Developed advanced Transformer, CNN, and multimodal deep learning models, enhancing accuracy in time-series forecasting, computer vision, NLP, and speech recognition applications
- Built end-to-end ML pipelines using PyTorch, TensorFlow, XGBoost, Flask, and Docker, emphasizing robust validation, generalization, and production-oriented experimentation
- Researched financial forecasting, multimodal perception, remote sensing, and large language model applications, contributing insights through international ML competitions and independent projects
Owner & Manager - French Press Café
(2013-11 - 2025-04)
Operated a profitable café for over 12 years, focusing on growth strategies
- Oversaw daily operations of a high-traffic café to ensure efficiency and customer satisfaction
- Led staff to consistently deliver high-quality customer service, enhancing customer satisfaction
- Oversaw finance operations, staffing, and strategic planning to maintain stability during market fluctuations
- Directed menu development and recipe innovation to improve customer experience and drive sales
- Cultivated decision-making intuition to improve operational efficiency and manage risks effectively