Job DescriptionJob Description
About the job
n
Mercor connects elite creative and technical talent with leading AI research labs. Headquartered in San Francisco, our investors include Benchmark, General Catalyst, Peter Thiel, Adam D'Angelo, Larry Summers, and Jack Dorsey.
n
Position: MLOps Engineer, LLM Performance (Kernels, Profiling, Serving)
nType: Contract
nCompensation: $90–$120/hour
nLocation: Remote
nCommitment: 40 hours/week
n
Role Responsibilities
n
nDesign challenging tasks across kernels, profiling, debugging, and inference serving. Write accurate, well-structured solutions to expose where frontier models fail on accelerator code.nGuide research and engineering teams to close knowledge gaps and improve AI model performance on ML systems, compilers, and training infrastructure topics.nEvaluate ML systems tasks and solutions. Provide clear written technical feedback that stands up to expert review.nDevelop guidelines and detailed rubrics covering kernel optimization, profiler output interpretation, distributed systems reasoning, and serving throughput and latency trade-offs.nCollaborate with other subject matter experts to keep training data consistent and accurate.nn
Qualifications
n
Must-Have
n
n2+ years of hands-on professional experience in ML systems, ML infrastructure, model serving, or accelerator performance engineering.nPractical experience in writing or optimizing custom kernels (Pallas, Triton, CUDA, CUTLASS),
performance profiling and trace analysis (JAX/XLA profiler, Kineto, torch.profiler, Nsight), debugging distributed or accelerator-bound workloads, serving large language models at scale (vLLM, SGLang, TensorRT-LLM, JetStream, KV cache, paged attention, continuous batching).nWorking production experience with JAX and/or PyTorch. Framework-level depth is a solid plus: custom operators, sharding and distributed training (pjit/shard_map, FSDP, Megatron, MaxText), or compiler and graph-level work (XLA/StableHLO, torch.compile, MLIR).nExperience with modern accelerators such as TPU (v4 to v6e), H100, B200, Trainium or MI300, and the ability to reason about throughput, latency, and memory trade-offs.nDemonstrable career progression.nAbility to engage reliably for at least 40 hours/week during weekdays.nStrong written communication skills and the ability to explain complex technical decisions clearly.nn
Compensation & Legal
n
nHourly contractornPaid weekly via Stripe Connectnn
Application Process (Takes 20–30 mins to complete)
n
nUpload resumenAI interview based on your resumenSubmit formnn
Resources & Support
n
nFor details about the interview process and platform information, please check: https://talent.docs.mercor.com/welcomenFor any help or support, reach out to:
[email protected]
PS: Our team reviews applications daily. Please complete your AI interview and application steps to be considered for this opportunity.
📌 MLOps Engineer - AI Systems (Canada)
🏢 Default Project
📍 Canada