Hello, I'm
Michael Cho

Ph.D. Candidate, Applied ML Scientist, Researcher, Engineer

Read the blog About me

Skills & Technologies

Technologies and tools I work with

Linux

Debian

Arch

Ubuntu

macOS

Windows Server

Android

iOS

Proxmox

Docker

Portainer

Kubernetes

Apptainer

AWS

EC2

PfSense

OPNSense

Tailscale

RDP

VNC

Nginx

Apache

EZDeploy

Datadog

PyTorch

Tensorflow

Hugging Face

Flower

OpenCV

OpenGL

CUDA

Spark

Dask

Pandas

Polars

Numpy

Numba

MATLAB

Jupyter

Slurm

HPC

Elastic

Snowflake

Outerbounds

Metaflow

MLflow

ONNX

Python

TypeScript

JavaScript

Node.js

C++

Java

Lua

HTML

CSS

Tailwind

Astro

MySQL

Git

Neovim

VSCode

Tmux

Zsh

Fish

Bash

Obsidian

Latex

Overleaf

Claude Code

Anaconda

Poetry

FastAPI

Pydantic

Pytest

Ruff

CircleCI

GitHub Actions

Jira

Confluence

News

Recent updates, milestones, and accepted publications

View all

2026.10

Cylindrical Geodesic Flow Matching for Quasiperiodic Physiological Signal Transformation

NeurIPS 2026 Main Conference, Sydney. This work introduces a flow matching method for paired cardiovascular waveform translation that replaces the standard affine path with a closed-form geodesic on the phase-amplitude cylinder, outperforming interpolation baselines on zero-shot photoplethysmography and limited-support seismocardiography adaptation benchmarks.

Paper

2026.08

Towards Monitoring Acute Opioid Withdrawal Severity Using a Multimodal Wearable Chest Patch

ACM Transactions on Computing for Healthcare. This study characterizes 32 cardiorespiratory features from wearable signals in 16 individuals with opioid use disorder over 7 days, identifying features that differ between high and low withdrawal and classifying withdrawal severity with machine learning.

Paper

2026.07

EffSCG: Real-Time Seismocardiogram Denoising on Edge Devices

IEEE Journal of Biomedical and Health Informatics. This work presents an optimization framework that lets a deep learning seismocardiogram denoiser run directly on resource-constrained wearables, cutting inference time and memory footprint without substantive accuracy loss.

Paper

2026.06

SCRIBE: Simultaneous Cardiac and Respiratory Impedance Biomarkers with Electromechanics

IEEE Sensors Journal. This work presents SCRIBE, a compact, flexible chest patch that concurrently measures cardiac and respiratory mechanics using ECG, seismocardiography, photoplethysmography, and bioimpedance.

Paper

View all news