AI Engineer, Model Quality and Performance
Cerebras · Sunnyvale, CA
About this role
## AI Engineer, Model Quality and Performance ### About Cerebras Cerebras Systems builds the world’s largest AI chip—56x larger than GPUs—enabling industry-leading training and inference speeds (over 10x faster than GPU-based hyperscale cloud inference). This performance leap is transforming AI application experiences by enabling real-time iteration and more agentic computation. Cerebras works with leading model labs, global enterprises, and AI-native startups. OpenAI recently announced a multi-year partnership with Cerebras to deploy 750 megawatts of scale: https://openai.com/index/cerebras-partnership/ --- ### About the Role You’ll own **model quality and performance** for Cerebras’ inference offerings. You’ll define what “good” looks like across the models we serve, build AI-driven systems to measure it at scale, and translate those signals into artifacts customers and product teams can actually use. You’ll use **AI agents** to create custom eval suites per customer use case, mine trajectories for representative test data, automate repetitive release qualification work, and help build performance datasets and benchmarking workflows. You’ll sit between **engineering, product, and customer-facing teams**. --- ### What You’ll Do - **Design eval suites with AI agents in the loop**: curate a thoughtful mix of advanced, basic, long-context, and customer-use-case-specific evals for every model release. Use Claude to generate, validate, and prune candidate test cases quickly. - **Build custom evals for target customers**: orchestrate AI agents to mine trajectories from customer workloads and synthesize representative eval sets. - **Automate eval execution end-to-end**: build AI-driven pipelines on top of standard tooling (Docker, Git, CI) so the system runs itself between releases—not a script you rerun manually. - **Forecast and benchmark performance**: build automations to predict and benchmark model performance on Cerebras for top customers, including how fast customer-specific workloads will run in production. - **Create product-quality tooling**: synthesize quality + performance data into a single, easy-to-use view. --- ### Skills & Qualifications - **Experience building AI agents**: you ship real systems with Claude (or equivalent) as a force multiplier—things that would be infeasible solo without agentic loops. - **Strong math/stats background**. - **Comfort with Docker, Git, and automation tooling**. - **A taste for tooling design**: you’ve shipped something non-engineers can use without complaining (bonus if AI helped you ship it). --- ### Assets (Nice to Have) - Performance-tuning experience on **custom silicon, GPUs, or FPGAs**. - Experience designing evals for **agentic / coding / long-context / multimodal** use cases. - Familiarity with open-source eval frameworks (e.g., **EvalScope, lm-eval-harness**). - Additional experience building AI agents. --- ### Why Join Cerebras People who are serious about software make their own hardware. With dozens of model releases and rapid growth, Cerebras is at an inflection point. Key reasons team members join: 1. Build a breakthrough AI platform beyond GPU constraints. 2. Publish and open source cutting-edge AI research. 3. Work on one of the fastest AI supercomputers in the world. 4. Enjoy job stability with startup vitality. 5. A simple, non-corporate culture that respects individual beliefs. Learn more: https://www.cerebras.ai/join-us --- ### Equal Opportunity Cerebras Sy
Listing freshness
CronJobs last confirmed this listing 2h ago. If its source stops confirming the opening for seven days, this page is removed from active inventory.