← All roles

ML Systems - Member of Technical Staff

Simile· Research
Apply on company site
San Francisco$200k - $400kFull-timePosted 2mo ago

About the role

Simile is building a production-focused ML systems platform to enable researchers to train, evaluate, and deploy large-scale simulations that model populations of agents. The role combines systems engineering, data pipelines, model serving, and optimization for society-scale simulations.

Responsibilities

  • Build the ML platform our researchers live in. Design and operate the services, libraries, and tooling that cover the full lifecycle - data exploration, feature generation, experiment tracking, training orchestration, evaluation, and deployment.
  • Make training and data pipelines fast. Own throughput end to end: model FLOPs utilization across training configs, tokenization cost when the data mix changes, and ingestion paths that take hours today where they should take minutes.
  • Make serving fast and cheap enough to run a society. Own the inference path our simulations run on: batching and scheduling, KV cache reuse across agents sharing context, quantization, and the request patterns unique to population-scale runs where one study is millions of interdependent calls.
  • Scaling simulation Data. Lead the redesign of our data architecture to handle the complexity and sheer volume of our simulation models.
  • Own the GPU cluster. Keep a multi-node fleet healthy and saturated: node bring-up, topology-aware NCCL and RDMA configuration, scheduling and queue depth, storage lifecycle and checkpoint capacity, autoscaling of serving capacity alongside training jobs, and the alerting that tells us when GPUs are sitting idle.
  • Engineer scientific evaluations. Build evaluation tooling with rigorous statistical frameworks that prove fidelity of simulations and run fast enough to be part of the development loop.
  • Push the state of the art. Reproduce, critique, and improve upon academic work in simulation, training, and inference optimization; translate breakthroughs into production improvements and document them with academic rigor.

Qualifications

  • Have deep systems and ML proficiency. High proficiency in Python and hands-on experience with modern ML frameworks (e.g., PyTorch, JAX).
  • Understand modern ML architectures well enough to optimize them. Comfortable around NVIDIA GPUs and the surrounding stack (NCCL, InfiniBand and NVLink topology, CUDA, Triton) or can get there quickly.
  • Have built production ML platforms or MLOps systems. Shipped infrastructure other engineers and researchers build on (training orchestration, experiment tooling, model serving, or LLM application platforms).
  • Have architected, observed, and debugged production distributed systems. Bonus if performance-critical and if rebuilt/refactored at scale.
  • Own the deployment pipeline. Understand the training and fine-tuning lifecycle and architect continuous ingestion, monitoring, and high-availability serving.
  • Are research- and data-literate. Navigate ML research frontier, reproduce papers, handle messy data, and document findings rigorously.
  • Are self-directed and pragmatic. Identify high-impact problems, learn quickly, and choose between ideal solutions and practical adjustments.
  • Have a humble attitude and strong collaboration mindset.
  • Have a strong quantitative foundation (CS/Math/Statistics or related) with demonstrated ability.

Skills

Required:PythonPyTorch / JAXGPU computing / NVIDIA stackDistributed systems / MLOpsDeployment & inference optimizationData engineering & pipelines

Benefits

  • Salary Range: $200,000 – $400,000 USD
  • Equity: Grants available for eligible roles, subject to board approval
  • Health & Wellness: Medical, dental, and vision coverage
  • Time Off: Flexible time off policies to support work-life balance
Last updated Oct 26, 2026. Interested? Apply on the company site.

About Simile

Company profileWebsite

More open roles at Simile