Portfolio at cat5.dev
Former full-stack platform service engineer for easy.gg
| larger projects | |
|---|---|
| 🦦 otterwatch | sea otter cam CV detection project |
| 📚 RAGical library (WIP) | cozy local RAG study tool for close reading and note taking |
| 🤟 openhand | ASL computer vision model training and inference on live video feed |
| 👩🚀 captionaut | video caption generation for content creators |
| 🚀 airship.top | live player count tracker |
| 🏎️ F1Guessr | f1-themed geoguessing game |
| 🎧 F1 Pit Wall | f1 driver radio transcription viewer |
| ☁️ Soundscape | FMA audio embedding visualizer |
| explorables | |
|---|---|
| 🦴 Backbone Benchmark Bracket (WIP) | benchmark of various CV backbone architectures using equivalent hyperparameters |
| 🍎 CV Segmentation Playground (WIP) | computer vision segmentation playground |
| 🔍 CV Detection Playground (WIP) | computer vision detection playground |
| 👀 ViT Playground (WIP) | interactive playground to explain vision transformer architecture |
| 🏁 Transformer Playground (WIP) | interactive playground to explain transformer architecture |
| 🔬 CV Interpretability Explorer | comparison of 5 CV interpretability methods across CNN/ResNet/ViT models |
| 🕰️ CNN Architecture Comparison | comparison of the major 6 CNN architectures of the past few decades |
| 🖼️ CNN Visualizer | demo of a pure PyTorch CNN trained on CIFAR |
| 🖍️ CNN Playground | entirely client-sided CNN single and multi-layer kernel playground |
| 🧩 BPE Playground | GPT-2 tokenization algorithm playground |
| python packages | |
|---|---|
| configurable, modular framework for live stream object detection | |
| 🎵 dinscribe | audio cleaning + transcription |
| 🙆 dinnote | audio cleaning + transcription + diarization |
| 💬 prompt2dataset | LLM to web query for rapid dataset generation |