AI & LLM Engineer

Job description

Job Summary Join the Digital Transformation Team to design, build, and maintain production-grade AI Agent systems that drive operational efficiency across Corning Taiwan's manufacturing plants. You will work alongside our existing AI Agent Engineer to scale the platform's agent ecosystem — evolving from individual agents toward an intelligent, autonomous multi-agent framework capable of auto-suggestion, self-reporting, and knowledge-driven decision support. Key Responsibilities Design and develop LLM-powered AI Agents using LangChain / LangGraph, including multi-agent architectures (Supervisor, Plan-Execute-Reflect loop, Sub- agent orchestration) Build and maintain RAG (Retrieval-Augmented Generation) pipelines forknowledge agents — document ingestion, vector search, and context management Develop Text-to-SQL capabilities with domain-specific schema understanding and query validation logic Engineer robust agent workflows with Pydantic structured output, idempotency design, and artifact-based data pipelines Maintain and improve accuracy of 7+ deployed production agents; triage edge cases and prompt failures Collaborate with process engineers and plant operations to translate domain knowledge into agent skills and tools Build agent evaluation frameworks: QA test cases, output quality monitoring, and continuous improvement loops Support the evolution of agent architecture from current Supervisor-Tool pattern toward Deep Agent design Experiences/Education BS/MS in Computer Science, Information Engineering, or related field 3-5 years of software or AI/ML development experience (strong project portfolio from new graduates welcome) Proficient in Python; experience with LLM application development Familiar with at least one LLM / Agent framework (LangChain, LlamaIndex, AutoGen, or equivalent) Understanding of RAG concepts: chunking, embedding, vector stores (FAISS, Chroma, etc.) Experience with SQL and relational databases • Comfortable working with ambiguous, real-world manufacturing data Preferred Qualifications: Experience with LangGraph, multi-agent orchestration, or Plan-Execute-Reflect patterns Familiarity with Pydantic for structured output and schema validation Experience deploying AI applications in production environments Knowledge of manufacturing processes or MES systems is a strong plus Good English proficiency in reading and writing for technical documentation