AI Engineer

Caterpillar Inc.•Irving, TX

13d•$97,530 - $158,480•Onsite

About The Position

The AI Engineer designs, develops, and optimizes advanced AI interactions for production-grade software applications. The role builds robust, scalable, and reliable systems using large language models (LLMs), orchestrates autonomous AI agents, implements Retrieval-Augmented Generation (RAG) pipelines, and establishes testing and observability frameworks. It bridges business logic and machine intelligence to deliver high-quality, safe, and deterministic AI behaviors. The position is onsite five days a week, with required locations in East Peoria, Illinois, or Irving, Texas. Caterpillar describes itself as a global manufacturer of construction and mining equipment, engines, turbines, and locomotives, supported by dealer networks and financing services.

Requirements

Expert-level mastery of structured prompt formats, including XML and JSON wrappers, system-level steering, and context-window optimization.
High proficiency in Python or TypeScript for backend AI logic, asynchronous data streams, and API integrations.
Hands-on experience building applications with LangChain, LangGraph, or native SDKs.
Solid experience with vector databases such as Azure AI Search or AWS OpenSearch and hybrid search retrieval methods.
Strong data manipulation capabilities using Pandas and NumPy, including data cleaning and token handling.
Foundational understanding of NLP transformers.
At least 2 years of dedicated professional experience developing production software integrated with GenAI/LLM pipelines.
A proven portfolio demonstrating deployed RAG systems or operational AI agents in commercial environments.

Nice To Haves

Prior implementation of observability tools such as Application Insights, OpenLLMetry, or custom evaluation and guardrail layers.
Knowledge of deep learning architectures, fine-tuning frameworks such as QLoRA and PEFT, and platforms such as Hugging Face, PyTorch, or TensorFlow.
Experience deploying containerized AI microservices using Docker, Kubernetes, AWS, or GCP.
Bachelor’s or master’s degree in computer science, Data Science, Artificial Intelligence, Computational Linguistics, Mathematics, or another highly technical discipline.

Responsibilities

Architect, version, and refine system, contextual, and dynamic prompts using techniques such as Chain-of-Thought, Tree-of-Thoughts, ReAct, and Few-Shot learning.
Reduce hallucinations, jailbreaking, and prompt injection risks through programmatic guardrails and adversarial prompting strategies.
Adapt, benchmark, and fine-tune prompt structures across open-source and proprietary models.
Design and orchestrate multi-agent systems and multi-step reasoning workflows using frameworks such as LangGraph and LangChain.
Build and optimize RAG pipelines, including semantic search, chunking strategies, embedding models, and vector database queries.
Implement function-calling and tool-use prompts that enable LLMs to interact with external APIs, databases, and legacy codebases.
Build automated testing suites with RAGAS and Evals frameworks to assess RAG accuracy and continuously validate and benchmark outputs.
Create regression testing suites and CI/CD pipelines to validate prompt stability, agent trajectories, and tool-calling accuracy.
Implement production LLM monitoring and observability using platforms such as Azure Application Insights and Microsoft Foundry.
Track and optimize semantic drift, token usage, cost efficiency, and end-to-end latency.
Work with Product Managers, Data Scientists, and Software Engineers to translate commercial requirements into deterministic AI capabilities.
Stay current with developments in LLMs, evaluation frameworks, and orchestration libraries.