PhD candidate researching artificial intelligence.
I turn messy questions into reproducible evidence — and product ideas into working software.
RESEARCH / 01
Reproducible AIS forecasting with strong baselines and CPA/TCPA risk evaluation. PYTHON · PUBLIC DATA · REPRODUCIBLE EVIDENCE |
RESEARCH / 02
Consistency screening for ship efficiency labels from public THETIS-MRV reports. PYTHON · SCIENTIFIC WRITING · OPEN RESEARCH |
PRODUCT / 03
A SwiftUI focus app using Screen Time rules, sessions, and reward tokens. SWIFTUI · FAMILYCONTROLS · PRODUCT DESIGN |
EVALUATE / 04
Coding-model evaluation across benchmarks, correctness, quality, and performance. LLM EVALUATION · BENCHMARKS · REPORTING |
// EXPERIMENT_LOG — smaller tools, prototypes, and learning projects
| Project | What I explored |
|---|---|
mcp_server_mysql_windows |
A local-first MySQL MCP service with parameterized queries and safe defaults |
ChatLingo |
Product concept and iOS architecture for AI-assisted English learning |
awkward-keyboard |
A focused language-spelling practice tool |
pptgpt |
An early experiment in generating a presentation from one sentence |
| Lens | What it means in practice |
|---|---|
| Research | Strong baselines, explicit splits, auditable evidence, measured claims |
| Product | Focused loops, useful defaults, working prototypes before feature volume |
| AI engineering | Evaluation before hype; agents as tools, not magic |
| Craft | Clear writing, calm interfaces, and systems that explain themselves |
RESEARCHING AI · BUILDING PRODUCTS · MAKING MUSIC · WRITING FICTION