Fast Volleyball Tracking Inference: Real-time volleyball ball detection and tracking at 100 FPS on CPU (Intel i5-10400F). Powered by an optimized ONNX model, outputs ball coordinates to CSV, with optional video visualization. Ideal for sports analytics and computer vision research.

A high-performance C/C++ inference server for Qwen3-ASR , optimized for CPU/GPU real-time streaming speech recognition.

An optimized C/C++ implementation of the HDBSCAN algorithm for the course Advanced Systems Lab.

French AI Law Assistant 🤖 — Local RAG chatbot for French AI regulations. Processes PDFs and delivers instant answers with Llama 3.2 1B via Ollama. Ultra-fast, no API needed.

Offline multimodal AI system with speech, vision, and language processing. LangGraph orchestration, Phi-4-mini LLM, Vosk STT, BLIP vision. Complete privacy, zero data transmission. Inspired by Interstellar's TARS.

Real-time, CPU-optimized face recognition pipeline (Webcam/RTSP) with FastAPI, InsightFace & ONNXRuntime. Includes live HUD & telemetry.

Nano-SLM: Extreme CPU-optimized Small Language Model with Tkinter Dashboard

Run a YOLO model over a folder of images to label the data, draw labelled bounding boxes, and save grouped detections to results.json. Configurable model, batch size, and resolution; CPU-friendly.

CPU-fast Redrob candidate ranking pipeline with structural trap detection, relevance filtering, and offline LLM coherence penalty.

Experimental neural network architecture with internal reasoning loops.