Darren Chien — SPARQNOW

At-a-Glance

Managing Director, APAC at Positron AI
Positron AI builds hardware and infrastructure for AI inference
Focuses on memory-first architecture for large models, long contexts, and decode-heavy serving
Leads the development of products like Atlas, an FPGA-based inference platform, and Asimov, a custom inference accelerator
Oversees the full inference system development, including silicon, servers, and software
Ensures an OpenAI-compatible runtime that supports standard Hugging Face model formats

Biography

Darren Chien is the Managing Director, APAC at Positron AI, where he is involved in building hardware and infrastructure for AI inference. Positron AI's approach addresses the constraints of modern inference, such as memory capacity, usable bandwidth, power, and cost, through a memory-first architecture.

This architecture is specifically designed for demanding workloads, including large models, long contexts, and decode-heavy serving.