Eli J a Hyeo Xu an-Y e

Technology & Automation Intern at Graticule Hedgie Fund

Surveyed and prepared reports on latest AI and technology developments, providing concise insights into emerging trends, breakthroughs, and potential applications. Developed a custom Python web scraping pipeline to extract real-time stock intelligence from Reddit (JSON endpoint scraping) and Yahoo Finance via API, plus other financial forums. Streamlined news updates and portfolio movements for portfolio managers using Google’s Gemini Flash 2.0 via API integrated with a Python auto-scheduler to summarize and curate financial news periodically, tailored to portfolio movements and stock watchlists.

Technology & Automation Intern at Graticule Hedg Fund

Surveyed and prepared reports on latest AI and technology developments, providing concise insights into emerging trends, breakthroughs, and potential applications. Developed a custom Python web-scraping pipeline to extract real-time stock intelligence from Reddit (via JSON endpoint scraping) and Yahoo Finance via API, as well as other financial forums. Streamlined news updates and portfolio movements for portfolio managers using Google’s Gemini Flash 2.0 via API integrated with a Python auto-scheduler to summarize and curate financial news periodically, tailored to portfolio movements and watchlists.

Center for Frontier AI Research (CFAR) Research Intern at A*STAR – Centre for Frontier AI Research (CFAR)

Conducted a comprehensive survey on Generative AI in Finance, assessing strengths and limitations of different models for various financial applications and producing a holistic landscape for finance professionals. Extracted key techniques from 200+ papers across GANs, VAEs, diffusion models, and LLMs, mapping their suitability to different financial applications. Authored the “Applications of Generative AI” section covering innovations across asset management, robo-advisory, credit risk, and regulatory automation. Provided benchmarking methods for GAI models in finance settings, highlighting gaps in domain-specific metrics.

Research Intern at A*STAR – Centre for Frontier AI Research (CFAR)

Conducted a comprehensive survey on Generative AI in Finance, assessing strengths and limitations of various models for different financial applications and providing a holistic landscape for financial professionals. Extracted key techniques from 200+ papers across GANs, VAEs, diffusion models, and LLMs, mapping suitability to different financial applications. Authored the “Applications of Generative AI” section covering innovations across asset management, robo-advisory, credit risk, and regulatory automation. Provided methods for benchmarking GAI models in financial settings, highlighting the gap in domain-specific metrics.

Institute of High Performance Computing (IHPC) Research Intern at A*STAR – Institute of High Performance Computing (IHPC)

Designed and fine-tuned a U-Net-based deep learning model in PyTorch for medical imaging, achieving over 85% segmentation accuracy for detecting retinal fluids in scans for early XLR S detection and prevention. Visualized segmentation outputs using Matplotlib and applied contrast-tuned augmentation for improved feature differentiation. Addressed detection accuracy bottlenecks by correcting pixelated fluid identification via data augmentation techniques and integrating advanced training methods (IoU-driven loss functions and Adam optimization) for robust learning. Partnered with Singapore General Hospital to validate the model’s utility in diagnosing XLR S progression, enabling early intervention tooling.

Computing Research Intern at A*STAR – Institute of High Performance Computing (IHPC)

Designed and fine-tuned a U-Net based deep learning model in PyTorch for medical imaging, achieving >85% segmentation accuracy for detecting retinal fluids in scans for early XLR S detection and prevention. Visualized segmentation outputs using Matplotlib and applied contrast-tuned augmentation for improved feature differentiation. Resolved bottlenecks in detection accuracy by correcting pixelated fluid identification using data augmentation techniques and integrating advanced training techniques (IoU-driven loss functions and Adam optimization) for robust learning. Partnered with Singapore General Hospital to validate the model’s utility in diagnosing XLR S progression, paving the way for early intervention tools.

Palliative Care Project Research Assistant at Tan Tock Seng Hospital – Palliative Care Project

Worked on research for palliative treatment to improve treatment support for patients. Compiled qualitative and quantitative insights into patient needs from doctor and patient palliative care interviews.