资深AI工程师 - (MCP & LLM集成)

This is a remote position.
Senior AI Engineer - ( MCP & LLM Integration)

About the RoleWe are looking for a high-performing Senior AI Engineer (including MCP) to lead the development and delivery of enterprise Model Context Protocol (MCP) integrations in Phase 1, and governed multi-model access, evaluation, and observability in Phase 2.

The ideal candidate will combine strong backend software engineering in Python with direct hands-on experience implementing the Model Context Protocol (MCP), LLM APIs (Anthropic Claude, OpenAI), tool-calling, and evaluation frameworks, ensuring robust, production-grade agentic workflows and reliable model interfaces across the engagement.

Programme Context & Phased Effort Allocation

The engagement is structured as a 22-week sequential programmedivided into two delivery phases, assuming a five-day working week:

• Phase 1: MCP Gateway and Governance (Weeks 1–10 / 10 Weeks):Full-time allocation (1.0 FTE) developing reference MCP servers, API adapters, central registry onboarding checks, and validating Claude client tool-use workflows.

• Phase 2: Model Access through Microsoft Foundry (Weeks 11–22 / 12 Weeks):Full-time allocation (1.0 FTE) integrating Claude, OpenAI, and specialist models, implementing streaming token protocols, usage attribution, and model evaluation harnesses.

Key Responsibilities

• Reference MCP Servers & Adapters:Develop reference MCP servers and API adapters, including JSON-schema tool definitions, granular permissions, input validation, and graceful error handling.

• Registry Integration & Onboarding:Implement MCP registry integration, automated onboarding checks, request/response logging, and telemetry ingestion.

• Client Tool Connectivity:Connect the selected Claude client to approved MCP tools and thoroughly validate representative agentic workflows and tool-calling execution.

• Multi-Model Provider Integration:Integrate Claude, OpenAI, and agreed specialist model endpoints through the approved Azure APIM gateway and Microsoft Foundry.

• Streaming, Attribution & Evaluation:Implement and test provider interfaces, streaming token responses, tool use, usage attribution, prompt caching, and model evaluation suites.

• Automated Testing & Documentation:Automate integration tests, resolve defects, and document reusable integration patterns and standard operating procedures (SOPs).

Required Skills & Experience

• 6–10+ years of software engineering experiencewith senior proficiency in Python (FastAPI, Pydantic, asyncio) or agreed backend integration languages.

• Model Context Protocol (MCP) & Agentic Workflows:Hands-on experience implementing the Model Context Protocol (MCP) specification, JSON-RPC, tool-calling schemas, and client-server integrations.

• Enterprise LLM Integration:Extensive experience integrating Anthropic Claude and OpenAI APIs, handling authentication, token rate limits, prompt caching, and streaming protocols.

• Model Evaluation & Observability:Experience with automated model evaluation, benchmark datasets, prompt engineering, tracing (LangSmith / OpenTelemetry), and cost/usage attribution.

• API Design & Automated Testing:Strong background in RESTful APIs, JSON schema validation, automated testing (pytest), mock services, and CI/CD integration.

• Security & Data Governance:Understanding of zero-trust API security, payload sanitization, credential handling, and enterprise data privacy.

Preferred Qualifications

• Capital Markets / Financial AI Applications:Experience deploying AI applications, document processing, or market research copilots in financial institutions.

• Microsoft Foundry & Azure AI:Experience with Microsoft Foundry, Azure OpenAI Service endpoints, and Azure APIM policies.

• Agent Frameworks:Familiarity with LangChain, LangGraph, AutoGen, or Semantic Kernel frameworks.

• Open Source Contributions:Active contributions to open-source AI tooling, MCP server repositories, or developer SDKs.

Key Competencies

• Model Context Protocol (MCP) Architecture

• Advanced Python Engineering & Asyncio

• LLM API Integration (Anthropic Claude & OpenAI)

• Agentic Tool-Calling & Schema Validation

• Model Evaluation, Benchmarking & Prompt Optimization

• Automated Testing & Integration Test Suites

• Production Telemetry & Token Usage Attribution

• Developer Tooling & Pattern Documentation

Requirements

Required Skills & Experience

• 6–10+ years of software engineering experiencewith senior proficiency in Python (FastAPI, Pydantic, asyncio) or agreed backend integration languages.

• Model Context Protocol (MCP) & Agentic Workflows:Hands-on experience implementing the Model Context Protocol (MCP) specification, JSON-RPC, tool-calling schemas, and client-server integrations.

• Enterprise LLM Integration:Extensive experience integrating Anthropic Claude and OpenAI APIs, handling authentication, token rate limits, prompt caching, and streaming protocols.

• Model Evaluation & Observability:Experience with automated model evaluation, benchmark datasets, prompt engineering, tracing (LangSmith / OpenTelemetry), and cost/usage attribution.

• API Design & Automated Testing:Strong background in RESTful APIs, JSON schema validation, automated testing (pytest), mock services, and CI/CD integration.

• Security & Data Governance:Understanding of zero-trust API security, payload sanitization, credential handling, and enterprise data privacy.

Benefits

Key Competencies

• Model Context Protocol (MCP) Architecture

• Advanced Python Engineering & Asyncio

• LLM API Integration (Anthropic Claude & OpenAI)

• Agentic Tool-Calling & Schema Validation

• Model Evaluation, Benchmarking & Prompt Optimization

• Automated Testing & Integration Test Suites

• Production Telemetry & Token Usage Attribution

• Developer Tooling & Pattern Documentation

Originally posted on Himalayas