I'm Pranshu, a Computer Science student from India who enjoys building things across the stack — from **AI/ML models **.
I'm especially interested in understanding how things work under the hood, rather than only using existing abstractions.
- 🔭 Currently building Setu — a native Android runtime for Windows
- 🧠 Working with AI/ML, Computer Vision & Deep Learning
- 🪟 Contributing to Microsoft PowerToys
- 📱 Building with Android & Flutter
- ⚙️ Interested in systems programming, runtimes & developer tools
- 🌐 Also build full-stack applications when the idea demands it
- 🚀 Exploring open source and looking for interesting problems to solve
"If it doesn't exist, build it. If it exists, figure out how it works."
A Wine-style native runtime designed to run Android APKs on Windows without a traditional virtual machine.
- Android → Windows compatibility layer
- Native runtime architecture
- Skia-based rendering
- Experimental systems-level project
- Built from the ground up
An AI-based deepfake detection pipeline using deep learning and computer vision.
Technologies:
PyTorch EfficientNet MobileNetV3 ASPP OpenCV Computer Vision
- Frame-based video analysis
- Hybrid CNN architecture
- Frame skipping optimization
- Exploring ONNX / TensorRT deployment
Systems Programming
Artificial Intelligence & Deep Learning
Windows Internals
Android Runtime Architecture
Open Source Development
Computer Vision
Compilers & Programming Languages
I'm actively interested in contributing to projects where I can work on real-world engineering problems, especially:
- Windows / developer tooling
- Systems & runtime development
- AI/ML
- Android
- Developer infrastructure
