docker

The Ultimate Developer Experience Platform for the Container Era

Exploring NLP weak supervision approaches to train text classification models. The project is also a prototype for a semi-automated text data labelling platform. Approaches: Snorkel and Zero-Shot Learning.

GraphRAG over any codebase — a Neo4j knowledge graph of code entities plus a 3072-dim vector index, bridged so one natural-language query walks structure and ranks by meaning. Python · Neo4j · OpenAI embeddings · Streamlit · Docker. Hybrid semantic + structural code search.

Open Voice OS Speech-to-Text (STT) container images and docker-compose.yml file for x86_64 CPU architecture.

A Streamlit-powered machine learning playground that automatically detects classification or regression tasks, builds pipelines with preprocessing, trains models interactively, and visualizes metrics using Plotly. Backward-compatible, fully responsive, and deployable on Streamlit Cloud or Docker.

Memcortex is a lightweight semantic memory layer for LLMs using open source embedding model and Weaviate for vector storage and retrieval.

OpenAI-compatible bridge that exposes Codex CLI/SDK as /v1/chat/completions with multimodal + JSON schema

PostgreSQL-native semantic search engine with multi-modal capabilities. Add AI-powered search to your existing database without separate vector databases, vendor fees, or complex setup. Features text + image search using CLIP embeddings, native SQL joins, and 10-minute Docker deployment.

Build Fullstack GenAI apps and AI Agents

Unraid CA template and Docker build for Mem0 (OpenMemory). This mega-container utilizes s6-overlay to seamlessly bundle the pre-compiled Qdrant Vector Database, the Python FastAPI/MCP server, and the Next.js Dashboard UI into a single, click-and-play local AI memory layer. Built for homelabs, privacy, and true agentic AI memory retention.

jupyter notebooks to fine tune whisper models on Vietnamese using Colab and/or Kaggle and/or AWS EC2

ORLA is a web application that transforms text prompts into detailed 3D models using advanced AI technologies. With an intuitive interface and powerful backend, ORLA enables users to generate high-quality 3D assets quickly and easily.

Local voice-to-Obsidian pipeline. Faster Whisper on CPU — no cloud, no API keys. Auto-triggered via systemd (Linux/macOS) or Docker.

Collection of recent advanced RAG techniques.

MLOps — vLLM + LiteLLM + Portainer CE + Netdata. Case study song ngữ: kiến trúc, metrics, bài học vận hành. · Self-hosted LLM inference done the MLOps way, a bilingual case study

Streamlit app for 3D semantic segmentation of intra-oral teeth scans with MeshSegNet

Cloud-agnostic Agentic AI & Multi-Modal RAG platform with GraphRAG, intelligent routing, MCP enterprise integrations (Odoo, PostgreSQL), and pluggable LLM providers.

Self-hosted Dakera AI memory server — Docker Compose, Kubernetes, Helm, HA cluster setup, and monitoring.

vLLM + Qwen3.6-27B (BF16) OpenAI-compatible inference server on AMD Strix Halo (Ryzen AI Max+ 395, gfx1151). Vision input, 256K context, /v1/responses with separated reasoning, via TheRock ROCm.

📊 ETL pipeline for YouTube competitor analytics. Orchestrated with Airflow, Docker, and AWS. Features sentiment analysis and a Power BI dashboard.