Senior Backend AI Engineer — Multi-Agent Orchestration

Socket.dev; San Francisco (CA); On-site; USD 173,000 - 260,000; Full time

Job summary

Salesforce is seeking a Trailblazer to join the Agentforce Reasoning Engine team as a production AI backend engineer. Based in San Francisco, CA or Bellevue, WA, you will build scalable services powering multi-agent orchestration, enabling agents to reason, plan, and act at scale.

You’ll collaborate with PMs, architects, and AI researchers, own stateful session management with Redis, and optimize performance across the stack, from design to production.

Qualifications

• 6+ years of experience building production AI systems and/or backend services.
• Strong Python programming skills and experience building production-grade async services (FastAPI or equivalent).
• Hands-on experience with distributed, scalable systems and modern data storage frameworks such as Redis, Kafka, and Docker/Kubernetes.
• Experience integrating with LLMs and prompt engineering in production systems.
• A related technical degree required.

Responsibilities

• Build and own scalable backend services powering multi-agent orchestration, supporting thousands of concurrent tenants at production scale.
• Design and improve stateful session management using Redis and distributed storage patterns, ensuring reliability across LLM call chains.
• Drive system performance through automation, including capacity planning, monitoring, and root cause analysis.
• Collaborate with Product Managers, Application Architects, and AI researchers to bring agent reasoning capabilities from concept to production.

Skills

Python

FastAPI

Async services

Redis

Kafka

Docker

Kubernetes

LLMs

Prompt engineering

Distributed systems

Education

Bachelor's degree in a related field

Tools

Redis

Kafka

Docker

Kubernetes

Job description

Salesforce is seeking a Trailblazer to join the Agentforce Reasoning Engine team as a production AI backend engineer. Based in San Francisco, CA or Bellevue, WA, you will build scalable services powering multi-agent orchestration, enabling agents to reason, plan, and act at scale.

You’ll collaborate with PMs, architects, and AI researchers, own stateful session management with Redis, and optimize performance across the stack, from design to production.