Deskripsi Pekerjaan
Synthetix Labs is scaling high-throughput, low-latency distributed inference engines for frontier foundation models. We are seeking a Staff AI Platform Engineer to lead core runtime architecture, maximize GPU efficiency across multi-region clusters, and build enterprise-grade serving infrastructure.
In this role, you will define the distributed systems backbone powering real-time inference across thousands of accelerators, driving reliability, performance optimization, and operational excellence.
Tanggung Jawab
- Architect and scale distributed GPU orchestration platforms for massive concurrent LLM inference workloads.
- Design low-latency model serving runtimes achieving sub-millisecond execution overhead at high load.
- Profile compute, memory, and network bottlenecks across NVIDIA accelerator clusters using CUDA profiling tools.
- Automate cluster provisioning, multi-cloud failover, and telemetry pipelines using infrastructure-as-code.
- Collaborate with ML research teams to package, benchmark, and deploy cutting-edge generative architectures.
- Establish distributed tracing, observability, and SLA benchmarks for critical enterprise platform services.
Kualifikasi
- 7+ years of professional backend engineering experience building distributed systems in Go, Rust, or C++.
- 3+ years deploying and operating mission-critical containerized services on production Kubernetes clusters.
- Hands-on experience with modern model-serving runtimes such as vLLM, TensorRT-LLM, or Triton Inference Server.
- Deep understanding of Linux systems internals, network protocols (gRPC, TCP, HTTP/2), and distributed storage systems.
- Demonstrated experience tuning accelerator performance and resolving low-level memory/compute bottlenecks.
- BS or MS in Computer Science, Computer Engineering, or equivalent production engineering experience.