AI Engineer from Vientiane, Laos 🌏 — I build multi-agent systems, AI-reliability tooling, and speech tech for low-resource languages, and I ship them end-to-end: model → API → deployed product.
- 🔭 Currently building phonalign — pip-installable audio-to-phoneme forced alignment for TTS datasets (English, Lao, ~100 languages)
- 🤖 I like agents that do real things — gate deployments, cross-examine alerts, block vulnerable code
- 🌱 SDG-focused builder — AI for recycling, smart cities, and Lao-language technology
- 📫 Reach me at danny.ssd7@gmail.com
| Project | What it does | |
|---|---|---|
| 🛡️ | DeployGuard | Multi-agent CI/CD security gate — five specialist AI agents collaboratively review every pull request and block vulnerable code before it ships. Live webhook, real deployment gating. Built for the Band of Agents Hackathon. |
| 🏛️ | DRIFT | Early-warning system for AI quality degradation — a prosecutor, defense, and judge cross-examine every drift alert with deterministic statistics before it fires. Live demo → Built for the AMD MI300X hackathon. |
| ♻️ | SARA | AI waste identifier and campus recycling intelligence — snap a photo, get material classification and disposal guidance; orgs get heatmaps, dashboards, and gamified engagement. Live app → |
| 🗣️ | phonalign | Audio-to-phoneme forced alignment for TTS dataset preprocessing — TextGrids, VITS filelists, exact duration arrays, and QA reports. English, Lao, and ~100 espeak languages with zero system dependencies. |
Focus areas: multi-agent systems · LLM fine-tuning · AI observability & drift detection · TTS / forced alignment · low-resource NLP (Lao)
Building AI that earns trust — from Vientiane to production. 🚀