Freshman Software Engineering Student @ Ajou University
I build things at the intersection of ML theory, systems, and real-world deployment.
I'm a 1st-year Software Engineering student who cares deeply about understanding why algorithms work, not just running them. My work spans ML research tools, physics simulation, NLP, and practical web services — all grounded in solid mathematical and engineering principles.
- 🔬 Core interests: Deep Learning Theory · Physics-Informed ML · Systems & OS Fundamentals
- 🏥 Applied goals: Medical AI (diagnostic imaging) · Industrial AI (process optimization)
- 📐 Philosophy: derive before you import; optimize before you scale; question before you trust
| Project | What it is | Stack |
|---|---|---|
| Gradient Pathology | Production-grade PyTorch gradient diagnostics library — heatmaps, Sankey flow diagrams, real-time Streamlit dashboard, and a 7-rule ExpertEngine with code-snippet recommendations. v0.8.0 | PyTorch Plotly Streamlit Python |
| Physics-Informed ML | End-to-end physics simulation platform using Fourier Neural Operators and PINNs. 100–1000× faster than traditional PDE solvers, with Bayesian UQ, FastAPI backend, Three.js 3D frontend, and Terraform IaC. | PyTorch FastAPI React Three.js Terraform |
| ML Gradient Descent Viz | MLP & backpropagation built from scratch with NumPy. Mathematical proofs, loss landscape visualization, and convergence analysis. | NumPy Python |
| Physics-Informed Optimizer | Optimization algorithms informed by physical priors. | Python SciPy |
| Chemical Reaction Rate Prediction | Compared Linear Regression, RF, and SVR for predicting reaction rates. Validated results against Arrhenius equation via feature importance. | Scikit-learn Python |
| Project | What it is | Stack |
|---|---|---|
| Ajou Dorm Finder | Full-featured web app for Ajou University students — eligibility checks, score calculation, interactive room-type charts, and D-Day countdown for the new dorm. Responsive across all devices. | React TypeScript Vite |
| AI Disease Classifier | End-to-end pneumonia detection web service from X-ray images. Solved class imbalance and reduced model size with ONNX quantization to fix cloud deployment crashes. | Flask ONNX JavaScript |
| Mini OS Simulator | Process scheduling algorithms and memory management logic implemented from scratch for OS fundamentals study. | Python C |
| NPU Simulator | Neural Processing Unit simulator exploring hardware-level inference acceleration concepts. | Python C |
| Project | What it is | Stack |
|---|---|---|
| AI Cyberbullying NLP Analysis | NLP pipeline for detecting and analyzing cyberbullying patterns in text. | Python NLP |
| Category | Technologies |
|---|---|
| Languages |     |
| ML & Science |     |
| Backend & APIs |    |
| Frontend |    |
| DevOps & Cloud |     |
| Period | Institution | Details |
|---|---|---|
| 2023 – 2025 | Eunhye High School | — |
| 2026 – | Ajou University | B.S. Software Engineering · AI · Industrial Engineering · Bio-Informatics |
I keep a daily learning log at sinsangwoo/TIL — notes on algorithms, ML theory, systems, and anything worth remembering.
Open to interesting problems. Always learning.