I'm a CS undergrad at KIIT building real systems at the intersection of machine learning, full-stack engineering, and computer vision.
I don't just study algorithms — I'm a Codeforces Expert (1704) and LeetCode Knight (1942, top 3.49% globally) with 723+ problems solved. I care about the gap between a working prototype and a production-quality system — and closing it.
- 🔭 Currently building Axon Studio — a browser-native neural network playground with live decision boundaries, gradient flow, and shareable experiments
- 🛠️ Actively working on MiniDB — a database engine from scratch in C++ (LSM-tree, compaction, WAL)
- 🌱 Deep-diving into systems programming, distributed storage, and LLM infrastructure
- 💬 Ask me about computer vision pipelines, TF.js, Next.js, or competitive programming
- 🤝 Open to SWE / ML engineering internships and open-source collaboration
- 📫 Reach me at ayushkardev@gmail.com
| Project | What it does | Stack |
|---|---|---|
| Axon Studio | Browser-native ML playground — build, train, visualize neural nets with no setup | Next.js · TF.js · D3.js · Supabase |
| Football Analysis | 4-stage CV pipeline — player tracking, team classification, tactical metrics from raw video | Python · YOLOv8 · OpenCV |
| Sign Language Detection | Real-time ASL classifier at 28 FPS on CPU using MediaPipe keypoints | Python · TensorFlow · OpenCV |
| J.A.R.V.I.S | Modular desktop voice assistant — voice commands, app control, web automation | Python · PyQt5 |
| Code-Craft | Multi-language online IDE with Monaco editor, real-time execution, JWT auth | React · Node.js · MongoDB |
- TauricResearch/TradingAgents — Fixed a misleading OpenRouter API key error on setup (PR #680); adding
fsspec-based cloud storage for generated reports (Issue #807)
