Software developer in Tunisia building Python and TypeScript tools for AI workflows, document search, automation, and developer productivity.
I like turning messy local workflows into useful software: searchable dashboards, PDF tooling, scraping systems, and small tools that make repeated work easier to reason about.
| Project | What it shows |
|---|---|
| sessions-sage | Local AI dashboard for OpenCode sessions: FastAPI, SQLite, HTMX, Ollama, searchable timelines, and daily digests. |
| pdf-fuzz | Local-first PDF search app with React, Django, Elasticsearch, Celery, RabbitMQ, Redis, and Docker Compose. |
| pdf-mcp | MCP server for PDF extraction, metadata, layout-aware parsing, search, and OCR through JSON-RPC. |
| dayflow | Personal planning app with Next.js, React, Tailwind CSS, Drizzle, and local SQLite. |
| mlh-demo | Hybrid web automation and scraping system using FastAPI, Celery, Redis, Selenium, and containerized workers. |
| three-library | React Three Fiber sandbox for WebGL, audio-reactive scenes, GLSL shaders, GLTF models, and interactive experiments. |
Languages: Python, TypeScript, JavaScript, Shell
Backend and automation: FastAPI, Django, Celery, Redis, RabbitMQ, Selenium, Docker
Frontend and interaction: React, Next.js, Vite, Tailwind CSS, HTMX, React Three Fiber, Three.js
Search, documents, and AI tooling: Elasticsearch, PyMuPDF, OCR workflows, MCP, Ollama
Databases and tools: SQLite, Drizzle, Git, GitHub Actions, Linux
- Building local AI and developer-productivity tools that keep data on the user's machine.
- Improving document workflows around PDF extraction, search, OCR, and retrieval.
- Exploring practical automation systems for repetitive browser and data-entry work.
- Website: hazembz.xyz
- Email: hazembenbz@gmail.com
- GitHub: @HazemBZ



