AI Engineers (Claude)

Title: AI Engineers (Claude)

Duration: Fulltime

Location: Remote

Job Description:

• Build reusable skills, package them as plug-ins, and compose them into governed AI applications that business lines use
• Configure and tune foundational models for use cases, including system and prompt design
• Build retrieval augmented generation, embeddings, and multi step workflows over governed enterprise data
• Take business-built AI prototypes and productionize them into reliable, governed applications
• Build simple to moderate skills, plug-ins, and MCP connections directly, and route builds requiring new infrastructure or deep integration to Enterprise Engineering, then govern and publish the result
• Stand up evaluation harnesses, golden set testing, guardrails, monitoring, and logging so output is consistent and reliable at enterprise scale
• Apply access control, least privilege, and audit to AI assets, partnering with Security and Cloud on identity and credentials
• Provide production support for the skills, plug-ins, and applications the team releases
• Document skills, applications, and technical decisions for handoff, audit, and maintenance.

Required Skills:

• Five or more years of software or automation engineering, including two or more years building and shipping applications backed by large language models to production
• Direct, recent, hands-on experience building and publishing Claude skills or plug-ins. Broad experience with large language models or frameworks, without shipped skills and plug-ins, is not sufficient. This is the requirement most submittals miss.
• Strong Python, REST APIs, and enterprise integration, with sound fundamentals in testing, version control, and monitoring
• Hands-on experience with foundational models: system and prompt design, skill and tool building, and function and tool calling
• Retrieval augmented generation, embeddings, and vector stores, including chunking and reranking
• Evaluation and monitoring for systems backed by large language models, with a working feel for cost, latency, and reliability tradeoffs