An independent developer focused on the intersection of atmospheric science, big data processing, and large language models (LLMs). Building bias-free predictive architectures to decode extreme weather events.

• jra3q-ai-forecaster: An open-source data integration tool that extracts 1D/3D vertical atmospheric profiles from JMA's JRA-3Q netCDF4 datasets for unbiased LLM objective forecasting. Fully deployed on Streamlit Cloud.

• Languages: Python (Pandas, NumPy, NetCDF4, SciPy)
• Frameworks & UI: Streamlit, PyTorch
• Methodologies: Bias-Free Prompt Engineering, Dimension Slicing, Convective Instability Detection
• Data Domains: Reanalysis Fields (JRA-3Q, ERA5), Isobaric Analysis

• GitHub Apps: JRA-3Q AI Forecaster Live Demo
• Technical Knowledge: Deeply experienced with Qiita and Zenn ecosystems for tracking enterprise-level software engineering trends.

"Pure numerical tensors carry the physical laws; text prompts carry the cognitive bias. Let the models compute the atmosphere raw."