Senior Platform Engineer

present of AI; San Francisco (CA); On-site; USD 150,000 - 210,000; Full time

Job summary

Coval, based in San Francisco, is seeking a senior backend engineer to own the platform that evaluates conversational AI. You will build and operate the backend that coordinates distributed workers, queues, orchestration, and the data layer moving conversations, audio, and results between stages.

A strong candidate has years of building and operating real systems in Python and AWS, works with managed services and Terraform, and has run workloads at scale with cost awareness and initiative on a

Qualifications

• Several years building and operating real systems in Python and AWS.
• Experience with managed services and Terraform.
• Experience running systems under load and cost-aware design.

Responsibilities

• Own backend machinery for large-scale evaluations of conversational AI.
• Manage distributed workers, queues, orchestration, and the data layer.
• Move conversations, audio, and results between stages with reliability and cost control.

Skills

Python

AWS

Terraform

Distributed queues

Cloud cost

Tools

AWS

Terraform

Distributed systems tooling

Job description

Engineering · San Francisco, CA, US · Full-time · Senior level

This role owns the backend machinery that runs large scale evaluations of conversational AI, including distributed workers, queues, orchestration, and the data layer that moves conversations, audio, and results between stages. A strong candidate has several years building and operating real systems in Python and AWS, works comfortably with managed services and Terraform, and has run something under load. The job also calls for cost judgment, since every model call and storage byte carries a price, and for the initiative to decide what to finish, fix, or defer on a small team.

How it fits the AI era

The platform exists to test AI agents before and after launch, so the work sits close to AI without being model research. Much of the engineering is ordinary distributed systems and data plumbing, but the volume of simulated conversations and model calls makes cost control and throughput central. Judgment about what to build versus buy, and how to keep unit economics sane as load grows, stays human.

How it builds AI experience

A person here would gain hands on exposure to how evaluation pipelines for voice and chat agents are built and run at scale, including the infrastructure behind LoRA training jobs. That builds practical knowledge of serving and scoring AI systems under real load. Someone preparing could strengthen this by working with queue based pipelines, cloud cost tracking, and the operational side of ML workloads.

Worth showing

• Python and AWS systems
• Terraform and managed services
• Distributed queues and orchestration
• Cloud cost and unit economics

This is our own read, written from the employer's posting and the company's Spotlight. The employer's posting is the source of truth for the role.

About Coval

Enterprise simulation and evaluation platform that stress-tests voice and chat AI agents before deployment, per the company.