Member of Technical Staff - Distributed Systems

Sail Research, San Francisco

What you will do

Design and implement high-performance schedulers (admission control, queuing, priority, fairness, preemption, bin packing).
Build global routing and traffic management (latency-aware dispatch, predictive autoscaling, failover strategies, cache-aware routing).
LLM-specific routing optimizations, e.g. KV caching that lets us trade memory for compute, across the cascading layers of GPU RAM, CPU RAM, and NVMe flash.
Build deep observability: we want to trace every millisecond of our systems, and catch failures early enough that we can make things right before customers even notice.

Strong distributed systems fundamentals (concurrency, networking, databases, queues, performance engineering).
Eagerness to work with agents coupled with a healthy dose of skepticism. We start by focusing our attention on the highest-level design decisions first, and try to debate and justify each of decisions. Then, we use our finite review time on the critical pieces of the system that we must build good mental models of to do good work.

Sail Research, San Francisco

Design and implement high-performance schedulers (admission control, queuing, priority, fairness, preemption, bin packing).
Build global routing and traffic management (latency-aware dispatch, predictive autoscaling, failover strategies, cache-aware routing).
LLM-specific routing optimizations, e.g. KV caching that lets us trade memory for compute, across the cascading layers of GPU RAM, CPU RAM, and NVMe flash.
Build deep observability: we want to trace every millisecond of our systems, and catch failures early enough that we can make things right before customers even notice.

Strong distributed systems fundamentals (concurrency, networking, databases, queues, performance engineering).
Eagerness to work with agents coupled with a healthy dose of skepticism. We start by focusing our attention on the highest-level design decisions first, and try to debate and justify each of decisions. Then, we use our finite review time on the critical pieces of the system that we must build good mental models of to do good work.