MLOps Engineer, LLM Systems (Serving, GPU Kernels, Profiling)
Mercor
About this opportunity
Mercor lists this role for a leading AI lab's GenAI team. It is AI model training and evaluation work for people who already do ML systems engineering: you write and assess MLOps and ML-systems tasks and solutions that generate high-quality training data for frontier models. It is a hands-on systems role, not an applied-modelling or data-science one.
What the work involves
Who is behind it
Required qualifications
- 2+ years of hands-on professional experience in ML systems, ML infrastructure, model serving, or GPU and accelerator performance engineering.
- Practical experience in at least one of: writing or optimising custom GPU kernels (CUDA, Triton, Pallas); profiling and trace analysis (Kineto, torch.profiler, Nsight, XLA or JAX profiler); debugging distributed or accelerator-bound workloads; serving large language models at scale (vLLM, SGLang, TensorRT-LLM, Ray Serve, KV cache, paged attention, continuous batching). More than one is a strong plus.
- Working production experience with JAX and/or PyTorch; framework-level depth (custom operators, FSDP/DDP/DeepSpeed/Megatron, compiler or graph-level work) is a strong plus.
- Familiarity with accelerators such as A100, H100, B200 or TPU, and the ability to reason about throughput, latency and memory trade-offs.
- A record of career progression, and strong written communication.
Skills
- mlops
- machine learning
- gpu kernel programming
- cuda
- triton
- performance profiling
- pytorch
- jax
- distributed training
- llm inference serving
- python
Eligibility requirements
- Mercor lists this for the United States, Canada and the United Kingdom. This site currently serves US and Canadian residents only.
- Able to commit 40 hours per week on weekdays, with no other engagements (Mercor states "no conflicts").
- Mercor states it cannot support H-1B or STEM OPT candidates at this time.
What to expect
- Full-time: 40 hours a week.
- **Check the terms before you accept:** this listing calls the role W-2 employment with Cincinnatus LLC in its overview but says "independent contractor" in its payment terms. Ask Mercor which applies to you.
- Mercor's terms section also says the work is remote on your own schedule, projects can be extended, shortened or ended early, and payment is weekly via Stripe or Wise.
- Pay listed by Mercor at $90–$120 per hour.
Application preparation guide
How to prepare: MLOps Engineer, LLM Systems (Serving, GPU Kernels, Profiling)
Before you apply
- Open the Mercor listing and read it in full. This summary may be out of date.
- Confirm you meet the location and hours requirements.
- Choose the one or two of the four areas you know best and prepare a concrete example of each (a kernel you optimised, a profile you diagnosed, a serving bottleneck you fixed).
- Name the specific tools you have used in production (for example CUDA, Triton, vLLM, torch.profiler) rather than listing categories.
- Keep an up-to-date CV or LinkedIn profile ready, and any work you are allowed to share.
What to expect
Mercor and the AI lab decide who is selected, and how. Applying through this page does not guarantee a response, an interview or an offer. Check the pay and engagement terms on the listing itself before you accept anything.
Staying safe
A genuine listing never asks you to pay to apply or to share banking or ID details up front.
Guide last updated Oct 4, 2026.
The provider makes every selection decision. Applying through this page does not guarantee a response, an offer, or any earnings. Details can change on the provider's side after we verify them, so confirm on their site.