Julian Niño

I’m a Junior AI Engineer with 2+ years of hands-on experience designing, building, and deploying Generative AI applications and production-ready backend APIs. I focus on turning unstructured data into reliable, queryable systems that work in real production environments.

I specialize in implementing RAG and Agentic RAG pipelines for document-based question answering, including projects for Colombian government institutions. I build end-to-end data processing workflows—document parsing, embedding generation, prompt engineering, and LLM integration—using models such as OpenAI and Mistral.

In my recent role, I implemented an end-to-end, production-oriented data processing pipeline in Python to transform unstructured PDF documents into structured, queryable data. I designed and managed vector-enabled relational layers, exposed vector search through REST APIs in TypeScript, and expanded systems into Agentic RAG architectures using OpenAI Agents libraries. I also run systematic evaluation (e.g., benchmarking chunking configurations and using RAGAs) to improve accuracy, grounding, and contextual relevance.

I also build robust ML and automation platforms for document-heavy workflows—ETL/CDC pipelines, evaluation and auditing pipelines, and knowledge-graph pipelines—while keeping engineering standards high. From orchestrators like ZenML to scalable processing with AWS S3/MinIO, Coiled, and RabbitMQ, I’m driven by responsibility, collaborative problem-solving, and continuous learning.