Senior/Staff Applied AI Engineer — Build Production AI Agents

Socket.dev; San Francisco (CA); On-site; USD 190,000 - 260,000; Full time

Benefits offered by this job

Health, dental, and vision insurance

401(k) plan

Team lunches and dinners in the office

Unlimited PTO

Job summary

Echelon is building the AI platform for Business Operations. We are seeking a senior engineer to design and ship production agents that turn enterprise context into reliable actions.

You will create agent loops, tool-use patterns, evaluation systems, and multi-agent coordination that work on real BizOps problems, not just demos. Work closely with Product Engineering, own failures, and raise architectural quality.

Qualifications

• 4+ years building production software with ML/LLM
• Shipped an agentic system with production feedback
• Strong Python and/or TypeScript engineering, APIs, tests, async
• Experience with tool calling, retrieval, and multi-step workflows
• Judgment on model usage vs deterministic code vs human checkpoint
• Debug nondeterministic behavior; measurable failure modes
• High agency and clear communication in changing requirements

Responsibilities

• Design and ship production agents that plan and call tools
• Build orchestration for agent swarms and task decomposition
• Create evaluation suites for success, factuality, latency, cost
• Improve context engineering across prompts and memory
• Build guardrails for permissions and data security
• Compare models with production traces and outcomes
• Collaborate with Product Engineering to implement BizOps workflows
• Own failures in production and prevent recurrence

Skills

LLM/ML

Production software

Python/TypeScript

APIs & tests

Observability

Tool calling

Multi-agent systems

Debugging nondeterminism

Tools

Vercel AI SDK

OpenAI

Anthropic

Mistral

AI Gateway

Braintrust

Job description

Echelon is building the AI platform for Business Operations. We are seeking a senior engineer to design and ship production agents that turn enterprise context into reliable actions.

You will create agent loops, tool-use patterns, evaluation systems, and multi-agent coordination that work on real BizOps problems, not just demos. Work closely with Product Engineering, own failures, and raise architectural quality.