AI Engineer II

About the role

The AI Center of Excellence (AI CoE) brings together AI Engineers and Data Scientists to research, prototype, and deliver production-grade AI systems. Our mission is to leverage cutting-edge Generative AI, agentic frameworks, and LLM-powered solutions to protect our customers' attack surfaces. We partner closely with Detection and Response teams, including our MDR service, to build intelligent AI systems that operate autonomously, reason across complex security data, and continuously evolve. We operate with a creative, iterative approach, building on 20+ years of threat analysis and a growing patent portfolio. We foster a collaborative environment focused on learning, mentorship, and practical impact. If you're early in your AI engineering career and excited to build real-world agentic and GenAI systems in cybersecurity - this is your opportunity. The technologies we use include: Python - core language for AI/ML engineering LLM/GenAI toolchains: LangChain, LangGraph, HuggingFace Transformers Agentic AI concepts: tool-calling, ReAct patterns, single and multi-agent workflows RAG pipelines: vector databases, embedding models, basic retrieval strategies AWS cloud ecosystem: Bedrock, SageMaker, Lambda, S3 LLM observability & evaluation: Langfuse, Promptfoo (foundational exposure) CI/CD for ML/LLM systems: GitHub Actions, Jenkins Monitoring: CloudWatch, dashboards Data science stack: scikit-learn, PyTorch, pandas, NumPy (supporting capabilities) Rapid7 is seeking an SE-II AI Engineer - Agentic & Generative AI to join our AI Center of Excellence as we expand our GenAI and agentic AI capabilities. This role is focused on building strong foundations in LLM engineering, contributing to agentic AI workflows, and growing into a well-rounded AI engineer in a production cybersecurity environment. This role is ideal for someone who is: Developing hands-on skills in LLM orchestration, prompt engineering, and RAG pipelines Gaining real-world experience with agentic AI patterns and AWS-based GenAI deployments Building data science and ML fundamentals to complement AI-native work In This Role, You Will Agentic AI & LLM Systems Contribute to building agentic AI workflows - tool-calling, basic agent loops, and LLM-driven automation under senior guidance Assist in developing and maintaining RAG pipelines - document ingestion, chunking, embedding, and retrieval Implement and iterate on prompt engineering - few-shot prompting, chain-of-thought, structured outputs Work with LangChain / LangGraph for LLM orchestration and chaining tasks Support LLM evaluation tasks - writing eval datasets, measuring output quality, running benchmarks Contribute to observability and monitoring of LLM systems - latency, token usage, output quality dashboards Deploy and test LLM-powered features on AWS Bedrock, Lambda, and SageMaker Participate in prompt versioning and LLM CI/CD pipelines under guidance of senior engineers Assist with guardrail implementation and output validation for production GenAI systems Learn and apply agentic AI patterns - ReAct, tool-use APIs, and structured output parsing ML & Data Science Work on data acquisition, cleaning, enrichment, and transformation for AI/ML pipelines Build and evaluate supervised ML models (classification, regression) for security use cases Apply unsupervised ML techniques such as clustering and anomaly detection Contribute to malware detection and user behavioral models under senior guidance Support model deployment on AWS SageMaker and monitor performance using established dashboards The Skills You'll Bring Core - Agentic AI & LLM (Required) 2-5 years of experience in AI/ML engineering or software engineering with AI focus Foundational hands-on experience with LangChain or similar LLM orchestration frameworks Familiarity with prompt engineering concepts and techniques Basic understanding of RAG pipelines - what they are, how retrieval works, and where they're applied Awareness of agentic AI patterns - tool-calling, agent loops, ReAct Exposure to LLM evaluation - understanding what good vs. bad LLM output looks like and how to measure it Working knowledge of AWS Bedrock and/or SageMaker for AI/ML workloads Strong Python skills and a learning-first mindset Data Science & ML Working proficiency with pandas, NumPy, scikit-learn Solid understanding of supervised and unsupervised ML, feature engineering, and model evaluation metrics Exposure to deep learning frameworks (PyTorch / TensorFlow) Basic familiarity with model explainability (SHAP, LIME) Advantageous Exposure to cybersecurity, malware datasets, or threat detection domains Hands-on experience with vector databases (FAISS, Pinecone, OpenSearch) Familiarity with LLM observability tools - Langfuse, Promptfoo, or similar Working knowledge of ML/LLM CI/CD pipelines Basic understanding of multi-agent orchestration concepts #LI-SG1 About Rapid7 At Rapid7, our vision is to create