Agentic AI Engineer

Hexion Careers

1d

About The Position

Lead Hexion’s enterprise Agentic AI and automation strategy—architecting and scaling agentic AI capabilities, identifying automation opportunities across all business functions, championing AI fluency and a learning culture, and overseeing the execution and KPI governance of automation initiatives to deliver measurable business value across the organization.

Requirements

• Bachelor’s degree in Computer Science, Engineering, Data Science, or a related technical field — or equivalent hands-on experience.
• 5+ years of professional software, AI/ML, or data engineering experience, including recent hands-on work with LLMs or agentic workflows (professional projects, internships, or substantive personal projects).
• Strong proficiency in Python (asynchronous programming, typing, FastAPI or similar) and solid software engineering fundamentals — APIs, Git, testing, and version control.
• Working knowledge of at least one agent framework: Microsoft AutoGen, Semantic Kernel, LangGraph, LangChain, LlamaIndex, or CrewAI.
• Practical experience with LLM APIs (Azure OpenAI, Anthropic Claude, OpenAI, or Hugging Face models) including prompt engineering, function/tool calling, and structured outputs.
• Understanding of core agentic concepts: multi-step reasoning, planning, tool use, memory, and multi-agent orchestration.
• Familiarity with RAG patterns, vector databases, embeddings, and retrieval evaluation.
• Exposure to cloud development on Microsoft Azure (Azure AI Foundry, Azure OpenAI, Azure Functions, Azure AI Search) or equivalent.
• Self-directed learner who stays current with a fast-moving field, reads primary sources, and can teach themselves new frameworks with minimal guidance.
• Strong problem-solving skills and the ability to work independently while collaborating across globally distributed teams.
• Clear written and verbal communication; able to explain technical concepts to non-technical stakeholders.

Nice To Haves

• Hands-on experience with Microsoft Copilot Studio, Microsoft 365 Copilot extensibility, or declarative agent plugins.
• Exposure to Palantir AIP, Foundry, or Ontology-driven AI applications.
• Experience with Databricks (Mosaic AI, Unity Catalog, MLflow) or Snowflake Cortex.
• Familiarity with Model Context Protocol (MCP), tool/function calling, and structured output schemas.
• Exposure to LLM observability and evaluation tooling — LangSmith, Langfuse, Arize, or Weights & Biases.
• Knowledge of GraphRAG, knowledge graphs, or semantic modeling.
• Experience with CI/CD (Azure DevOps or GitHub Actions), Docker, and infrastructure-as-code (Terraform or Bicep).
• Public portfolio — GitHub projects, technical blog posts, open-source contributions, or hackathon work — demonstrating curiosity and hands-on learning in generative or agentic AI.
• Prior experience delivering AI applications in an enterprise setting, particularly in process industries, manufacturing, supply chain, or commercial analytics.

Responsibilities

• Design and build agentic AI systems using modern frameworks such as Microsoft AutoGen, Semantic Kernel, LangGraph, LangChain, or CrewAI — including single-agent, multi-agent, supervisor, and handoff orchestration patterns.
• Develop end-to-end agent logic: planning loops, tool use, structured outputs, short- and long-term memory, state management, and human-in-the-loop checkpoints.
• Integrate Large Language Models (Azure OpenAI, Anthropic Claude, open-weight models) into business applications with attention to quality, latency, and cost.
• Build Retrieval-Augmented Generation (RAG) pipelines — document parsing, chunking, embeddings, hybrid search, and reranking — against vector stores such as Azure AI Search, Databricks Vector Search, Pinecone, or FAISS.
• Develop tools and connectors (including Model Context Protocol servers) that expose enterprise data sources and APIs as safe, well-documented actions for agents to call.
• Apply prompt engineering best practices and build evaluation harnesses — golden datasets, LLM-as-judge, deterministic assertions — to measure and regression-test agent behavior.
• Contribute to deployment and observability: package agents for production on Azure AI Foundry or Container Apps, add tracing (LangSmith, Langfuse, or OpenTelemetry), and help monitor token spend, latency, and failure modes.
• Implement guardrails for content safety, PII handling, and safe-action confirmations before any write operation against enterprise systems.
• Collaborate with product owners and subject-matter experts to translate business workflows into agent execution plans; prototype rapidly, demo, and iterate.
• Maintain high code quality through Git-based workflows, code reviews, unit and integration testing, and clear technical documentation

Benefits

• We invest in innovation, sustainability, and continuous development—equipping you with the tools, training, and opportunities to excel.
• With an unwavering commitment to safety, partnership, belonging, and impact, we empower you to lead change and strengthen industries worldwide.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior
• Number of Employees: 1,001-5,000 employees