I'm a systems engineer focused on bare-metal performance and on-device intelligence. I bridge low-level systems programming with modern AI—building backends in Rust and Spring Boot, native Android layers in Kotlin/C++ via JNI, and pushing local LLM inference to the limit with llama.cpp.
- 📱 Android systems — JNI bridges, geofencing automation, and secure runtime frameworks
- 🦀 Rust & C/C++ for zero-overhead backend engines and native Android modules
- 🤖 On-device AI — local LLMs, agentic workflows, and MCP (Model Context Protocol)
Sub-millisecond real-time interaction analysis via Rust ↔ Kotlin JNI bridge
Stack: Rust (Backend), Kotlin (Android), JNI, C++, Unit & Integration Testing
Highlights
- Bare-metal performance with sub-millisecond latency
- Rust ↔ Kotlin JNI bridge for memory safety and execution speed
- REST-inspired local communication protocol between the Rust engine and Kotlin frontend
- Extensive testing of native bridge components to prevent memory leaks
Geofencing-powered automation using the Android Geofencing API and Rhino JavaScript Engine
Stack: Kotlin, Android Geofencing API, Rhino Engine, Room (SQLite), XML
Highlights
- Research paper: "AutoFlow: A Secure Automation Framework for Android" (Machine Intelligence Conference)
- Trigger-action automation based on location, Wi-Fi, and Bluetooth states
- Sandboxed Rhino JavaScript execution environment for secure user scripting
- Local persistence using Room (SQLite) for automation triggers and execution history
Offline TTS engine converting PDF/ePub into natural speech with zero network dependency
Stack: Kotlin, On-Device TTS, NLP, Media3, Retrofit, Clean Architecture
Highlights
- Offline text-to-speech engine eliminating network latency
- NLP-based chapter boundary detection and contextual bookmark generation
- Reading progress synchronization using Retrofit-based REST APIs
- Clean Architecture separating media playback logic from the UI
Agentic workflows, MCP implementation, and local LLM inference
Stack: Rust, C++, llama.cpp, Custom MCP Servers
Highlights
- Autonomous agents running entirely on-device
- Local LLM inference with zero cloud dependency
- Model Context Protocol (MCP) implementation
- 🤖 On-device AI and Agentic Architectures
- 🦀 High-performance backend development with Rust and C/C++
- ☕ Scalable backend services using Spring Boot and Docker
- 📱 Android systems programming, JNI, and local LLM inference
