Top Tech Jobs & Startup Jobs in Toronto

The Deployment Engineer will manage AI inference clusters, optimizing deployment, capacity allocation, and ensuring reliability of pipeline operations across datacenters.

The role involves developing infrastructure for benchmarking AI inference performance, analyzing system behavior, designing telemetry systems, and collaborating with teams to improve feature performance.

Manage AI compute clusters, monitor systems health, optimize resources, and troubleshoot technical issues, ensuring high performance for ML applications.

You will prototype and benchmark model innovations, develop automation for experiments, and work with teams on software/hardware integration.

The Performance Engineer - Inference will optimize model inference speed and throughput, debug low-level kernel performance, and develop tools to visualize performance data.

Lead post-silicon bring-up and optimization of wafer-scale engines: develop debug flows, HW-SW codesign optimizations, build large-scale test infrastructure, create self-checking metrics and instrumentation, and collaborate across silicon, performance, and software teams to improve production performance and release processes.

Design and implement system-level debugging, validation, and observability platforms. Build automated anomaly detection, visualization and analysis tools, and frameworks for failure classification and regression detection. Extend compilers, runtimes, and instrumentation for advanced profiling. Improve bring-up and low-level debug workflows, partner cross-functionally across hardware, firmware, compiler and runtime teams, lead high-impact initiatives, and support incident response and long-term corrective actions.

Lead post-silicon bring-up and performance optimizations for Cerebras Wafer Scale Engines. Develop/debug production-ready flows, instrumentation, and self-checking metrics. Build infrastructure for large-scale silicon workload testing, collaborate with silicon architects, performance and software teams, and streamline CI/CD and release workflows.

Lead a team to improve the reliability of advanced compute clusters and manage failure analysis and debugging processes.

Join the SOTA Training Platform team to bring up ML models on Cerebras systems, enhancing performance across the software stack and debugging issues.