Nov 2023 – Present · 2 yrs 6 mos · Germany · On-site

• Researching curriculum learning strategies in reinforcement learning
• Focusing on enhancing agent robustness and overall performance
• Developing methodologies for creating ever-improving AI agents
• Aiming to enable agents to adapt to increasingly complex tasks and environments
• Exploring the intersection of curriculum design and reinforcement learning algorithms complex tasks and environments.

Aug 2022 – Oct 2023 · 1 yr 2 mos · Hsinchu City, Taiwan, Taiwan · On-site

• Optimized chip placement strategies utilizing deep reinforcement learning techniques
• Developed advanced thermal simulation models integrating U-Net architecture and physics-informed neural networks (PINN)
• Implemented and managed large language model fine-tuning services, leveraging proprietary data for enhanced performance

Sep 2020 – Aug 2022 · 1 yr 11 mos · Taipei, Taipei City, Taiwan · On-site

• Research on automatic music generation using Transformer-based models
• Developed variable-length music score infilling techniques with XLNet and specialized encoding
• Contributed to MidiBERT-Piano for large-scale symbolic music understanding
• Applied advanced NLP techniques to music generation and analysis