I’m Sai, an Applied AI and GenAI engineer based in Thailand. I turn ambiguous operational problems into systems that can be tested, measured, and used.
My work sits where agentic AI, computer vision, edge deployment, and Geospatial AI meet real constraints.
How I work
I approach projects with a Forward Deployed Engineer mindset: understand the actual workflow, build the smallest useful experiment, measure what changes, and keep the people operating the system in the loop.
That approach shaped my master’s research at the Asian Institute of Technology, where I built an oil-palm detection and canopy-measurement pipeline across multiple UAV image resolutions. The resulting AI Teacher Agent reduced manual annotation time by more than 80%.
What I write about
This blog is my public notebook for technical field reports, reproducible experiments, and lessons from building AI systems. I write about agents, model economics, computer vision, research workflows, edge AI, and the engineering gap between a promising demo and dependable infrastructure.
Some essays live here; others are published on Medium. Both are part of the same body of technical work and appear together in the writing archive.
Away from the screen
I’m also an IPF-registered competitive powerlifter. Training has taught me to value patient iteration, honest measurement, and showing up when progress is not immediately visible.