🚀 Benchmark deep learning engines across mobile and desktop platforms with KMPMLBench for effective performance evaluation using Kotlin Multiplatform.

Automate MVP builds by running an AI agent that picks goals and delivers working solutions without user intervention.

Verify Edge AI model binaries securely using wolfSSL with zero dynamic memory, optimized for embedded and constrained environments.

🔌 Run Edge TPU models on Linux with a pure-Python USB engine, no libedgetpu or tflite_runtime, supporting multiple architectures.

Chinese Chess game on Android

Mondrian: Edge-Enabled High-Performance Video Analytics with Compressive Mixed Inference

🧬 High-performance TensorFlow Lite library for React Native with GPU acceleration

fully OSS lightweight network video recorder system witten in C with modern js frontend

Converted TFLite Model Zoo for Android with LiteRT CompiledModel GPU acceleration

A repository for storing models that have been inter-converted between various frameworks. Supported frameworks are TensorFlow, PyTorch, ONNX, OpenVINO, TFJS, TFTRT, TensorFlowLite (Float32/16/INT8), EdgeTPU, CoreML.

A offline-first mobile application designed to support elderly individuals with emergency assistance, medication management, and well-being monitoring.

High-performance browser-based AI library for real-time object detection (YOLO12) and pose estimation (RTMW) using ONNX Runtime Web

A hardware implementation of quantized neural network inference in the PyRTL hardware description language.

TFLite image classification of dogs

11ms NPU object detection on i.MX 8M Plus | MIPI camera → TFLite INT8 → HDMI | Kernel drivers, device tree, WiFi/BT, Yocto BSP — full-stack embedded Linux + edge AI

Hardware-aware face detection on Samsung GT-S7392 (ARM Cortex-A9)

DPDFNet: causal single-channel speech enhancement that boosts DeepFilterNet2 with dual-path RNN blocks for stronger long-range temporal and cross-band modeling. Repo includes PyTorch implementation + checkpoints, ONNX & TFLite models with inference code, and a real-time demo.

Real-time Object Detection Lab | Android NDK & TFLite | Jetpack Compose | MVI & Clean Arch | Coroutines & Flow | Hilt | Macrobenchmark. A showcase of scalable and performant Android Engineering.