Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|
Monitor • Diagnose • Optimize • Benchmark • Learn
GPU Insight AI is an open-source Android application designed for AI engineers, machine learning practitioners, DevOps teams, researchers, and students. It combines GPU monitoring, AI-assisted diagnostics, benchmarking, infrastructure insights, and educational tools into a modern mobile experience.
Built with Kotlin • Jetpack Compose • Material 3 • Gemini AI • Room • Hilt • MVVM • WorkManager
Modern AI workloads depend heavily on GPUs, but understanding GPU performance, memory bottlenecks, CUDA issues, and infrastructure health often requires multiple tools.
GPU Insight AI brings monitoring, diagnostics, AI-powered explanations, and performance insights together in one Android application.
Whether you're training large language models, debugging CUDA issues, managing GPU infrastructure, or learning GPU architecture, GPU Insight AI aims to make complex GPU systems easier to understand.
- AI-assisted GPU performance analysis
- CUDA & PyTorch error explanation
- Out-of-memory (OOM) troubleshooting
- NCCL communication diagnostics
- TensorFlow & JAX log analysis
- Intelligent optimization recommendations
- Real-time GPU utilization
- VRAM usage monitoring
- Power consumption
- Temperature tracking
- Fan speed monitoring
- GPU clock frequencies
- Driver information
- Hardware health insights
- NVLink topology visualization
- Multi-GPU monitoring
- MIG partition visualization
- PCIe bandwidth monitoring
- ECC error detection
- Tensor Core utilization
- Triton Inference Server health checks
- GPU specification database
- Historical metrics
- Interactive charts
- GPU utilization trends
- Thermal analysis
- Performance benchmarking
- Resource usage reports
- Performance scorecards
Gemini AI provides guidance for:
- CUDA optimization
- Batch size recommendations
- Mixed precision (FP16/BF16)
- Memory optimization
- Performance bottleneck analysis
- Training optimization
- Inference optimization
- Privacy-first AI requests
- Sensitive data redaction
- Role-Based Access Control (RBAC)
- Audit logging
- Secure local storage
- Android Keystore integration
Generate professional reports including:
- GPU health reports
- Performance summaries
- Benchmark reports
- Markdown export
- PDF export
- Executive summaries
Interactive educational content covering:
- GPU Architecture
- CUDA Fundamentals
- Tensor Cores
- VRAM Management
- PCIe & NVLink
- AI Infrastructure
- Performance Optimization
| Category | Technologies |
|---|---|
| Language | Kotlin |
| UI | Jetpack Compose, Material Design 3 |
| Architecture | MVVM, Clean Architecture |
| Dependency Injection | Hilt |
| Local Storage | Room, DataStore |
| Networking | OkHttp |
| AI | Google Gemini API |
| Background Tasks | WorkManager |
| State Management | StateFlow |
| Build System | Gradle Kotlin DSL |
| CI/CD | GitHub Actions |
GPU Insight AI follows modern Android development practices.
Presentation Layer │ ├── Jetpack Compose ├── ViewModels ├── StateFlow └── Navigation ↓ Domain Layer Use Cases Business Logic Repository Interfaces ↓ Data Layer Room Network DataStore Repositories ↓ Infrastructure Gemini API Background Workers Notifications Storage
- Modern Android Architecture
- Clean Architecture
- MVVM
- Material Design 3
- Kotlin Coroutines
- StateFlow
- Hilt Dependency Injection
- Room Database
- WorkManager
- Google Gemini AI
- Modular Project Structure
- GitHub Actions CI/CD
- Apache 2.0 Licensed
- Open Source
git clone https://github.com/karthikrshet/GPU-Insight-AI.git
cd GPU-Insight-AI
cp .env.example local.properties
# Add your Gemini API Key
./gradlew assembleDebug
./gradlew test- AI Diagnostics
- GPU Monitoring
- Performance Benchmarking
- NVIDIA GPU Support
- AI Recommendations
- Plugin System
- Desktop Companion
- Cloud Synchronization
- Wear OS Companion
- Compose Multiplatform
Contributions are welcome.
Please read:
- CONTRIBUTING.md
- CODE_OF_CONDUCT.md
- SECURITY.md
before submitting issues or pull requests.
Licensed under the Apache License 2.0.
Copyright © 2026 Karthik Rajesh Shet (@karthikrshet)
⭐ If you find this project useful, consider giving it a star on GitHub.
Building open-source tools for AI engineers, developers, and the GPU computing community.