Ryan Maroof

I’m an AI Product Engineer Intern building responsible AI workflows and LLM engineering for cybersecurity use cases. I focus on designing evaluation pipelines that reduce misuse risk and enforce safe, auditable response handling.

At GetPostLabs, I apply Claude, prompt engineering, and risk classification to architect evaluation pipelines and product documentation, working with an internal team to develop LLM-based review logic for request screening, response safety checks, and structured audit trails. I also earned the Anthropic Foundations certification on Skilljar, completing hands-on tasks covering AI fluency, Constitutional AI, prompt engineering, and responsible AI principles.

My engineering work includes NyayBot, an AI Legal Assistant that builds a RAG-based chatbot with LangChain and FAISS—indexing 893 sections and enforcing strictly database-grounded answers with zero hallucination. I implemented multi-query retrieval to expand each user query into legal-terminology variants, and built an 8-script pipeline to parse and structure legal PDFs into JSON.

I complement this with applied ML and NLP through projects like NxtWatch, where I engineered a multi-domain recommender using cosine similarity matrices and an NLP feature pipeline across movies, books, and anime—then integrated TMDB and OMDb APIs for interactive recommendations. I’ve also proven my ML fundamentals by ranking in the top 2.4% of the Amazon ML Challenge 2025.