Senior, Agentic AI Engineer

American Bureau of Shipping•Houston, TX

About The Position

The Senior AI Agentic Engineer designs, builds, and operationalizes intelligent agent systems that automate complex enterprise business processes end-to-end. This role works at the intersection of LLMs, systems engineering, and applied machine learning — architecting multi-agent pipelines, tool-augmented reasoning systems, and retrieval-augmented generation (RAG) workflows across a range of enterprise platforms (e.g., Databricks AgentBricks, Azure OpenAI) and open-source frameworks (e.g., LangChain, LangGraph, AutoGen) — with the expectation that the right candidate brings familiarity with the broader and rapidly evolving ecosystem. The ideal candidate brings deep hands-on engineering experience with a proven track record of delivering agentic AI systems into production at enterprise scale — not just prototypes — applying rigorous software engineering principles, including modular system design, testability, resilience engineering, and security-by-design to ensure agents are maintainable, reliable, and safe in the long run. This means architecting for failure — building in retries, fallbacks, and graceful degradation — and treating latency and cost as first-class engineering constraints from day one, not afterthoughts discovered in production. Beyond technical delivery, the Senior AI Agentic Engineer mentors engineers across the team, shapes the organization's AI automation strategy, translates ambiguous business problems into well-structured agentic solutions, and drives the responsible and secure deployment of AI agents across business-critical functions.

Requirements

3 years’ work experience as an AI Agentic Engineer and over 7 years' experience in Data Science, Gen AI, Information Systems, Computer Science, Software Engineering, or other relevant field with relevant experience.
Required – bachelor’s degree in data science, Information Systems, Computer Science, Engineering, or other relevant field with relevant experience.
Technical — Agentic Frameworks & LLMs: Proven enterprise experience architecting and deploying production-grade multi-agent AI systems that automate real business workflows end-to-end — not just proofs of concept.
Deep hands-on expertise with agent orchestration frameworks such as LangChain, LangGraph, AutoGen, Semantic Kernel, DSPy, CrewAI, and platform-native solutions such as Databricks AgentBricks / Mosaic AI Agent Framework — with openness to emerging tools in the rapidly evolving ecosystem.
Deep understanding of LLMs and foundation models (e.g., GPT, Claude, Llama, Mistral, Gemini), including their capabilities, limitations, and appropriate use case fit.
Experience with structured outputs, function/tool calling, JSON schema design, and multi-turn agent loop engineering.
Technical — RAG & Data Platforms: Strong knowledge of RAG architectures, vector databases (e.g., Pinecone, Weaviate, Chroma, pgvector), embedding models, and hybrid retrieval strategies.
Hands-on experience with Databricks, including Unity Catalog, MLflow, Delta Lake, and Databricks Workflows for end-to-end data and AI pipelines.
Experience with database technologies, data lakes, and enterprise data platforms, including SQL, cloud storage, and streaming data sources that agents consume at runtime
Technical — Deployment, MLOps & Engineering: Strong Python proficiency and experience building production-grade services, APIs, and microservices that support agentic systems.
Experience deploying LLM and agent workloads on cloud platforms (e.g., Azure OpenAI Service, AWS Bedrock, Google Vertex AI) with containerized infrastructure (Docker, Kubernetes).
Experience implementing LLMOps practices, including experiment tracking (e.g., MLflow, W&B), prompt versioning, evaluation harnesses, latency profiling, and CI/CD for AI systems.
Experience with enterprise security, data governance, and compliance requirements for AI deployments, including PII handling, role-based access control, and audit logging.
Evaluation & Responsible AI: Familiarity with LLM evaluation techniques, failure mode analysis, red-teaming, and benchmark construction to maintain quality and trust in production agents.
Working knowledge of responsible AI principles, including fairness, explainability, safety guardrails, and human oversight mechanisms in agentic deployments.
Leadership & Communication: Strong written and verbal communication skills with the ability to explain complex GenAI and agentic concepts clearly to both technical teams and executive stakeholders.
Demonstrated ability to lead cross-functional AI projects from discovery through production, aligning engineering, data, product, legal, and business operations teams.
Ability to mentor junior team members, establish engineering standards, and foster a culture of experimentation and responsible AI development.
Ability to translate ambiguous, open-ended business challenges into structured agentic solution designs with clear scope and success criteria.
Display an entrepreneurial mindset with a bias for practical, high-impact solutions; comfortable operating in ambiguous environments and rapidly evolving technology landscapes.
Working knowledge of ABS Health, Safety, Quality, and Environmental Management System.

Nice To Haves

Preferred - Master’s Degree in Data Science, Data Analytics, Information Systems, Computer Science, Engineering, or other relevant field.
Prefer to have Gen AI Certifications

Responsibilities

Agentic AI System Design & Development: Design, build, and deploy end-to-end agentic AI systems using LLMs, tools, memory, and planning frameworks to automate complex, multi-step enterprise business processes.
Architect and implement both single-agent and multi-agent workflows for autonomous task execution, decision support, and orchestration — defining agent roles, memory strategies, tool integrations, and handoff protocols.
Develop tool-using agents with function calling, structured outputs, API integrations, database connectors, RPA hooks, and enterprise workflow triggers.
Lead the integration of agentic solutions with enterprise systems, including ERP, Business apps, and orchestration platforms such as Databricks, Airflow, and Azure Data Factory.
Retrieval-Augmented Generation (RAG): Design and optimize RAG pipelines, including document ingestion, chunking strategies, embedding models, vector store selection, and retrieval ranking for enterprise knowledge bases.
Implement advanced retrieval techniques such as hybrid search, metadata filtering, re-ranking, and query rewriting to improve grounding and reduce hallucination.
Evaluate and continuously tune RAG systems for accuracy, latency, factual grounding, and cost efficiency.
Model Adaptation & Prompt Engineering: Evaluate and select frontier and open-source LLMs (e.g., GPT-4o, Claude, Llama, Mistral, Gemini) and apply fine-tuning strategies — including instruction tuning appropriate to each business use case.
Optimize prompts, system instructions, and output schemas for reliability, determinism, and safety across agentic pipelines.
Apply reinforcement or feedback-driven optimization where applicable, including human-in-the-loop and automated evaluation loops.
Evaluation, Monitoring & Governance: Define evaluation frameworks for agentic systems covering task success, factuality, grounding, latency, cost, and failure mode analysis.
Build observability and monitoring pipelines for agent behavior, tool call traces, and runtime failure detection.
Partner with governance, risk, and compliance teams to ensure responsible AI practices, audit traceability, data privacy, and regulatory adherence across all deployed agents.
Production Deployment & LLMOps: Deploy GenAI and agentic systems into production using cloud-native architectures on platforms such as Azure, AWS Bedrock, or Google Vertex AI with containerized (Docker/Kubernetes) delivery.
Implement CI/CD pipelines, prompt versioning, rollback strategies, and runtime safeguards for LLM applications in enterprise environments.
Optimize deployed systems for performance, cost efficiency, and scalability under real-world load.
Collaboration, Mentorship & Strategy: Collaborate with software engineers, product managers, data scientists, and business stakeholders to translate ambiguous process challenges into well-structured agentic solutions.
Mentor AI engineers and data scientists on agentic design patterns, responsible AI practices, and production-grade engineering standards.
Contribute to the organization's AI automation strategy, co-authoring technical roadmaps, governance policies, and center-of-excellence standards for agentic AI.
Stay at the forefront of the agentic AI landscape, rapidly evaluating new frameworks and research findings and communicating their business relevance to leadership.