Software Engineer, Tinker Infrastructure
Core
Design, build, and operate distributed systems for job scheduling, resource orchestration, checkpointing, and fault recovery across large multi-node GPU clusters to power a post-training platform.
Role type
Senior IC distributed systems engineer (GPU infrastructure)
Builds
Distributed training and sampling infrastructure for a post-training platform
Domain
AI infrastructure / GPU clusters
Deliverable
infrastructure
Required skills
distributed systems design, consensus models, failure detection, job scheduling, state management, Python/Go/Rust/C++, incident response, observability, architectural tradeoffs
Preferred skills
large-scale GPU cluster operations, ML training/inference stack, checkpointing, distributed storage, high-throughput networking (RDMA/NCCL), developer-facing APIs
Technologies
Python, Go, Rust, C++, RDMA, NCCL
Responsibilities
Design and operate distributed systems for job scheduling and resource orchestration; ensure reliability and throughput of live production systems; handle node failures and network partitions; improve observability with metrics and tracing; lead incident response and root-cause analysis; partner with research teams to evolve platform architecture; mentor engineers on distributed systems practices
Seniority
Senior, hands-on IC with significant ownership