SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Simile is a simulation company that uses AI to model human behavior at scale, helping enterprises make better decisions. The company has grown rapidly since launching five months ago, achieving 5x revenue growth and securing $200M in funding at a $2B valuation from top-tier investors including Greenoaks and Index Ventures. Simile's foundation model has run tens of millions of simulations for Fortune 100 companies like CVS Health, Wealthfront, Deloitte, and Gallup.
The Research Infrastructure team builds the systems underlying the entire model lifecycle: data ingestion, distributed training, evaluation, serving, and monitoring. This role is unique because the research-to-product pipeline is exceptionally tight—experimental methods validated on Monday are integrated into systems customers use for high-stakes business decisions. Additionally, simulating a society requires running inference over populations of interdependent agents, not single requests. Cost per simulation and latency per agent directly determine what research is feasible to run.
As a Member of Technical Staff in Research Infrastructure, you will own the platform researchers train, evaluate, and deploy on—from experimental design through production serving. You will design data schemas and training pipelines, profile and optimize serving paths to reduce FLOPs and GPU memory usage, and manage the full deployment pipeline. Your work compounds across every researcher and simulation the company runs.
Key responsibilities include: building the ML platform covering the full lifecycle (data exploration, feature generation, experiment tracking, training orchestration, evaluation, deployment); optimizing training and data pipelines for throughput; designing inference paths for population-scale simulations with millions of interdependent calls; leading data architecture redesign to handle simulation complexity and volume; managing a multi-node GPU cluster across multiple compute providers; and building rigorous statistical evaluation frameworks that integrate into the development loop rather than serving as end-stage gates.
This role is ideal for engineers who find satisfaction in pushing their work to absolute limits and owning problems end-to-end, including the often-overlooked last mile of deployment. You will work directly with researchers and see immediate impact on what research becomes possible.
