One sophomore, two peer-reviewed papers, five research fronts

What it does This is a portfolio index, not a monolithic app. It collects five distinct research threads—two already peer-reviewed (IEEE AIAM 2025 and IJHSR 2026) and three in flight—spanning quantum-inspired genetic algorithms fused with graph neural networks, C. elegans mitochondrial genomics, pelvic-floor ultrasound kinematics at Stanford, a wildfire-response RL agent trained on LA terrain, and an empirical study on how biology diagrams mislead students. Each sub-directory contains its own code, manuscripts, and data pipelines.

The interesting bit Most high-school repos are tutorial forks; this one contains actual published papers with DOIs, an IEEE Xplore entry, and a live Cesium 3D demo of an actor-critic firefighting agent grounded in the January 2025 Palisades Fire. The author also documents honest limitations—synthetic traffic and unsourced constants in the RL project—rather than polishing them away.

Key highlights

• Lead-author IEEE paper on a quantum-inspired GA + GNN ensemble for multimodal classification, with SHAP interpretability and full synthetic data.
• IJHSR publication on C. elegans mitochondrial resilience using CaeNDR variants, Ensembl VEP, dN/dS tests, and a neural-net expression predictor reporting Spearman ρ = 0.959.
• InfernoTactics: an actor-critic wildfire RL agent (CNN + MLP) dispatching resources on real SRTM elevation and OSM data, with a live 3D demo and documented limitations.
• Stanford Medicine urology project analyzing perineal ultrasound kinematics (displacement → velocity → acceleration) across 23 subjects.
• Pre-data-collection study targeting CBE—Life Sciences Education on how simplified biology diagrams create student misconceptions.

Caveats

• This repo is largely a curated map; the heavy code and data live in sub-repositories, so you will need to dig past the main README for implementations.
• The Stanford pelvic-floor imaging and the CBE-LSE diagram study are preliminary or pre-data-collection, so there are no reproducible findings there yet.
• The IEEE model’s reported accuracy (0.53) beats the baselines listed but is still modest; the README frames it correctly as a proof-of-concept hybrid.

Verdict Worth browsing if you mentor early researchers, hire for breadth, or simply need a reminder that age and impact are not correlated. Skip it if you are looking for a single, production-ready tool to drop into a pipeline.