Role: AI Integration Engineer – Context Sources (MCP)
Location: Alpharetta, GA or New York, NY(Hybrid)
Role Summary
We are seeking an AI Integration Engineer to accelerate the onboarding and integration of AI development tools with enterprise context sources using Model Context Protocol (MCP) and similar frameworks. This role focuses on enabling developers to leverage AI tools effectively by connecting them with relevant internal and external data sources.
Key Responsibilities
- Design and implement integrations using Model Context Protocol (MCP) to connect AI tools with enterprise systems
- Build, validate, and maintain data pipelines between:
- Code repositories
- Wikis (Confluence, SharePoint)
- Ticketing systems (Jira)
- Internal APIs
- Evaluate and improve context quality, relevance, and usability for AI-assisted development
- Collaborate with security and data governance teams for:
- Access control
- Data classification
- Redaction policies
- Develop reusable integration patterns and documentation for production handover
- Work closely with developers to optimize AI tool adoption and performance
- Required Skills & Experience
- 2+ years of hands-on experience with AI-assisted development tools
- Strong programming skills in Python and/or TypeScript (Node.js)
- Experience building and consuming REST / GraphQL APIs
- Understanding of:
- Authentication (OAuth, tokens)
- Pagination & rate limiting
- Knowledge of Model Context Protocol (MCP) or similar LLM context-injection approaches
- Familiarity with:
- Vector databases
- Embeddings
- Search techniques (BM25, semantic retrieval)
- Preferred Qualifications
- Experience with RAG (Retrieval-Augmented Generation) architectures
- Knowledge of LangChain, LangGraph, or similar frameworks
- Exposure to enterprise tools:
- Confluence
- SharePoint
- GitHub Enterprise / Bitbucket
- Jira
- Experience in financial services domain
- Understanding of knowledge graphs or hybrid retrieval systems
- Education
- Bachelor’s degree in Computer Science, Engineering, or related field
