Engineering that starts with measurement

I'm Surapat Ek-In — most people call me Arm. I founded ArmLab to help organizations build AI systems that actually work at scale. Not prototypes that impress in demos, but production systems that survive real-world traffic, real-world data, and real-world constraints.

My background spans performance engineering, systems optimization, and applied research. I've spent years working at the intersection of hardware and software — understanding why systems slow down, where bottlenecks hide, and how to fix them without adding unnecessary complexity.

Before ArmLab, I worked on large-scale ML infrastructure, GPU-accelerated computing, and distributed systems. I hold a PhD and approach every problem the same way: measure first, hypothesize, validate, then optimize. No guessing. No cargo-cult engineering.

ArmLab exists because I kept seeing the same patterns: teams with brilliant ML researchers building systems that couldn't handle production load. Inference pipelines 10x slower than they needed to be. Training infrastructure that wasted compute. These are engineering problems, and they have engineering solutions.

I write about what I learn, share what works, and consult for teams that want to ship reliable AI products. If you're building something ambitious and want an engineering partner who cares about the details — let's talk.