Scientific computing with real-world impact

Optical Device Simulation & Analysis

Simulated photonic device behavior and optical coupling characteristics using Lumerical and MATLAB-based workflows.

• Conducted parameter sweeps across geometric variables such as gap spacing and waveguide widths
• Generated surface plots and statistical summaries from simulation outputs
• Interpreted results to inform design trade-offs and photonic performance optimization

Based on academic research simulation workflows in photonic device modeling.

Alan Watts Conversational AI

Designed and implemented a conversational AI system combining curated philosophical datasets with retrieval-augmented generation to balance stylistic fidelity and grounded responses.

• Dataset design and transformation for dialogic training data
• Retrieval and grounding strategy for style plus substance
• Deployment-oriented roadmap for a polished public demo

Focused on combining stylistic language modeling with grounded retrieval.

View repository →

Optical Diffraction Tomography Simulation

Simulation-driven photonic modeling pipeline demonstrating forward diffraction sampling and inverse reconstruction, highlighting how physical systems can be translated into computational inference problems.

View repository →

Health Knowledge RAG System

Designed a retrieval-augmented generation (RAG) system using public health data sources to generate grounded, context-aware responses to medical and wellness queries.

• Built document ingestion and chunking pipeline for structured medical content
• Implemented embedding-based retrieval using vector similarity search
• Integrated LLM generation with retrieved context to reduce hallucination
• Explored tradeoffs between model size, latency, and response quality

Built as part of a broader private backend experimentation environment.