I'm a physician and computer science undergraduate exploring how software engineering and artificial intelligence can advance radiological practice. My radiology experience is hospital-based diagnostic imaging, including emergency/trauma and oncologic care. I have hands-on contributions to DICOM viewers, PACS infrastructure, and exploratory AI tooling for healthcare and medical imaging, and I'm comfortable moving from clinical questions to code and back.
Main interests:
- Local-first imaging platforms and DICOM viewers (2D/MPR/volumetric rendering, DIMSE/DICOMweb)
- Generative AI in medicine and AI-assisted radiological decision-making (LLMs/agents, local inference, RAG/finetuning)
- Infrastructure for healthcare systems (PACS, NAS, EHR, and other systems)
- AI-assisted software engineering and developer tooling
- Full-stack development tailored to clinical workflows
Isis DICOM Viewer Cross-platform DICOM viewer family — for all Apple devices, Windows and Android |
dicom-slides Interactive DICOM viewer for PowerPoint and HTML slides — 2D stack, MPR and 3D. |
dicom-plugin · dicom-skill · dicom-viewer-mcp-app DICOM toolkit for AI agents: query/retrieve, send, anonymize, transcode, export, and an interactive in-chat viewer. |
TotalSegmentator OsiriX/Horos Plugin Run TotalSegmentator easily in OsiriX and Horos. |
CodeAtlas Explore your codebase from individual symbols to system-wide architecture. |
TaskForge Smart git tools, project planning and task workflows grounded in your codebase. |
The Legend of Source - Branches of Time Git history explorer with evidence-linked narrative. |
Scout Desktop editor for radiology reports: smart autocomplete tools, AI agents and speech-to-text. |
DICOM-Swift Pure Swift DICOM decoder toolkit for iOS/macOS — metadata and pixel parsing, window/level and export pipelines. |
MTK — Metal Toolkit Swift Package for Metal-native medical volume rendering — GPU-resident volume, MPR, projections and transfer functions. |
RadRef Modular collaborative radiology reference app. |
Podcast Intelligence Podcast and audio platform: transcription, time-aware summaries, and grounded Q&A with citations. |
Turing-Lab Touch-first cross-platform reimplementation of JFLAP: automata, grammars, pushdown automata, Turing machines, and regexes. |
BIONG Biologically Inspired Organism, Not a Game: a simulator where modular cortical networks drive survival behavior. |
MacPGP OpenPGP for macOS: key management, file/message encryption, signing and Keychain passphrase storage. |
Video Subtitler Local video dubbing app: transcription, translation, and dubbing. |
- Go-PACS — Native Go/Fyne PACS desktop app built on dicom-go — archive import, remote node management, C-ECHO/C-FIND/C-MOVE/C-STORE, and a built-in receiver.
- dicom-go — Pure Go DICOM module: read/write, DICOM JSON conversion, pixel data extraction, dictionaries and transfer syntaxes, and minimal C-ECHO/C-STORE workflows.
- rusty-dicom-node — Terminal-first Rust DICOM node client built on dicom-rs: remote node configuration, C-FIND/C-MOVE/C-STORE workflows, SQLite local indexing, recursive/ZIP import, and a ratatui TUI.
- PACS-Natural-Language-Query — CLI toolkit that translates natural language or SQL into validated DICOM C-FIND queries, with dry-run/explain output, result streaming, and PHI-aware logging.
- Dicom-Tools — Multi-language DICOM workbench comparing DCMTK, GDCM, ITK/VTK, dicom-rs, fo-dicom, dcm4che, Cornerstone3D, pydicom/pynetdicom, and related backends through a shared UI.
- dicompressor — Native C++20 batch DICOM transcoder for HTJ2K/JPEG 2000 lossless workflows, built with DCMTK and OpenJPH.
- OsiriX-Backup-Plugin — Sends DICOM studies to remote PACS/storage targets with queueing, retry, SHA-256 integrity manifests, and duplicate avoidance.
- orthanc-tools — Python operations toolkit for Orthanc: Docker/Ubuntu deployment, REST health checks, remote sync, PACS mirroring, backfill, and backup workflows.
- Orthanc for QNAP — Custom
.qpkgpackaging of the Orthanc server for QNAP NAS devices. - Radiology-Templates — Radiology report templates with Python/Rust tools for DOCX/Markdown/TXT round trips, formatting rules, and index generation.
- mammography-pipelines-py — Research package and CLI for breast-density experiments: preprocessing, feature engineering, reproducible reporting, and model training.
- brain-mri-pipelines-py — Alzheimer's MRI research framework: multi-view MRI + clinical-feature fusion, EfficientNet/DenseNet/MedicalNet backbones, and subject-level split safeguards.
- AutoComp — macOS menu bar autocomplete that watches accessible text fields and shows inline suggestions, backed by OpenAI-compatible endpoints, Apple Intelligence, or local llama.cpp.
- ReadMyPaper-go — Local app that converts scientific PDFs into listenable audio: Docling layout extraction, reading-order repair, optional LLM cleanup, notation verbalization, and Piper/Kokoro TTS.
- reports-to-llm — Rust preprocessing tool that normalizes DOCX/RTF medical reports into clean, structured UTF-8 text for LLM/RAG research.
- Rust-Neural-Networks · Swift-Neural-Networks — Educational from-scratch neural networks: MNIST, CIFAR-10, XOR, manual backprop, CNN/MLP/self-attention, plus MLX-based Swift workflows.
- hermes-local-patchkeeper — Hermes Agent skills for auditing, capturing, and safely reapplying intentional local customizations after updates.
- PGP-Client-go — OpenPGP client in Go: key generation/import/export, file and message encryption, signing, and signature verification.
- Code-Scanner — Rust CLI (with Bash fallback) that bundles a repository into a single text report for audits, code review, and LLM prompts.
- GitHub-Replicant — Async Rust CLI for bulk GitHub backup and sync of owned, starred, watched, and discovered repositories.
- jff-to-tex Turing Machine Diagram Converter — Python converter that turns JFLAP
.jffTuring machine diagrams into LaTeX/TikZ. - LaTeX-Paper-Template — SBC-based LaTeX skeleton for papers, monographs, and theses, with centralized metadata and release checklist.
- BTC-21M-Countdown — Live Bitcoin supply countdown to the 21 Million milestone.




















