Senior FA Engineer | Python & AI Enthusiast | Building scalable systems at the edge of LLMs and Computer Vision
I bridge the gap between complex AI models and production-ready applications — designing systems that are secure, efficient, and genuinely useful in the real world.
- 🔭 Senior FA Engineer & Process Engineer for digital watch manufacturing at Cal-Comp Electronics, Thailand
- 🧠 Deep interest in LLM inference, computer vision, and turning research-grade AI into deployable tools
- 🛠️ I like solving problems I actually run into at work — most of my projects started as "I needed this and couldn't find it"
- 🌱 Currently exploring self-hosted LLM infrastructure and privacy-first AI tooling for enterprise environments
- 📫 Reach me on LinkedIn
| Category | Technologies |
|---|---|
| Languages | C, Rust, Python, JavaScript |
| AI / ML | PyTorch, OpenCV, Scikit-Learn, Ultralytics YOLO |
| Backend | FastAPI, Flask, Django |
| Database / Cloud | PostgreSQL, MongoDB, SQLite, CockroachDB, Docker, GCP |
| Embedded Systems | C, MicroPython |
inferX_llm_stack A self-hosted, GPU-accelerated LLM inference stack running behind HTTPS. Delivers a ChatGPT-like web interface powered by your own models via vLLM, secured with SSL through Nginx, and backed by PostgreSQL for persistent storage.
spreadsheet_agent A private spreadsheet assistant built for teams that can't upload sensitive data to third-party AI platforms. Brings AI-assisted spreadsheet work in-house, without compromising company data policies.
File-Sharing-Web-App A lightweight Streamlit app for sharing files via randomly generated access codes — no email threads or direct links required.
EDS_Database_Management A secure data collection application for managing Energy Dispersive X-ray Spectroscopy (EDS) analysis results, integrated with VHX microscopy images.
Ultralytics_YOLO_Object_Detection_Testing_GUI A PyQt5 desktop application for real-time object detection using YOLO. Load a model, tune detection settings, and run inference on media through a clean GUI.
Label_Craft A Tkinter-based tool to create, edit, and export bounding box annotations in YOLO format — built to speed up dataset preparation for object detection training.
"Code is like humor. When you have to explain it, it's bad."