Senior Staff Ai Platform Engineer

📍 United States

USD 184,000-253,000 per year

SENIOR

✅ Remote

Tech Stack
Tag name is followed by "@" symbol and proficiency level value. About proficiency levels:

1-2 — basic awareness. Minimal hands-on experience, and a rudimentary understanding of the technology's purpose;

3-6 — daily use. Comfortable and regular usage, capable of handling common tasks and challenges related to the technology;

7-9 — you are an expert, you can teach others, you know all the pitfalls and tricks;

10 — exceptional knowledge, comprehensive understanding, and adeptness in all aspects of the technology, including advanced problem-solving. Think twice before claiming or demanding such level.

AI @ 1 API @ 4 AWS @ 4 Audit @ 4 FastAPI @ 7 LLM @ 1 Machine Learning Observability @ 4 Python @ 7 React @ 3 Security TypeScript @ 3

Details

Responsibilities

Design, build, and operate components of the AI Gateway, including:
- Centralized identity and authentication/authorization
- Token budgeting
- DLP and content guardrails
- Multi-model routing and failover
- MCP allow-listing
- Request-level audit logging
Build and maintain pieces of the Harness layer, including:
- Agent orchestration (LangGraph or equivalent)
- Prompt construction and context management
- Memory and state handling across multi-turn interactions
- Model abstraction across providers such as AWS Bedrock and Google Vertex
Develop and extend production services in Python (FastAPI) and pydantic.ai for LLM-powered components, and contribute to the React/TypeScript surfaces that expose platform capabilities to internal teams.
Implement MCP/tool wiring for internal and SaaS-embedded agents so every caller passes through the same policy controls.
Contribute to the Claude and Gemini Enterprise plugin framework by building, testing, and hardening role-based skills and agents, and help mature the pipeline for how plugins are authored, evaluated, and deployed.
Instrument the platform for observability including metrics, tracing, and audit trails, and participate in on-call and operational support for platform services.
Write clear technical documentation and participate in architecture and code reviews, maintaining a high bar for quality, security, and maintainability.
Partner with engineers across Enterprise Data, Enterprise Apps, Product Development, and Infosec to integrate the platform with governed data sources and shared architectural contracts.
Work with the Sr. Director and model evaluation tooling to close the loop between evaluation results and model selection, prompt tuning, and routing decisions.

Requirements

Hands-on experience building production services that sit in front of multiple consumers such as an API gateway, internal platform, or data platform, with direct exposure to authentication/authorization, rate limiting, observability, or audit logging. Direct AI and LLM platform experience is a strong plus but not required.
Practical experience with agent orchestration frameworks (LangGraph or equivalent) and calling hosted model providers such as AWS Bedrock or Google Vertex, with a working understanding of how context windows, memory, and tool-calling behave in production.
8 or more years of professional software engineering experience, with experience building or operating AI and ML infrastructure in a production environment.
Strong Python skills, ideally with FastAPI, and comfort picking up frameworks like pydantic.ai for LLM-powered components. Working familiarity with React and TypeScript is a plus.
Exposure to or curiosity about the Model Context Protocol (MCP) and the challenges of governing tool access for agents operating across enterprise systems.
Familiarity with enterprise AI deployments such as Claude Enterprise (Anthropic) or Gemini Enterprise is a plus, including how they are administered and how access and policy controls work.
Solid software engineering fundamentals including clean, tested, production-grade code, strong API design skills, and the ability to give and receive feedback well in code and architecture reviews.
Comfort operating in a fast-moving, still-forming platform environment where you are energized by ambiguity and enjoy turning a rough architecture into working, reliable infrastructure.