Luis Fernando Martinez

Development of Generative AI-based systems and process automation to solve various business problems. These services are exposed through REST APIs developed in Python with FastAPI and in Java with Spring Boot.

Document Extractor: Developed with high-precision LLM+OCR (Azure Document Intelligence) allowing structuring and extracting information from documents with 98% accuracy.

Advanced design and configuration of RAG systems for information retrieval and generation optimization, implementation of vector databases in private environments via Docker.

Chatbot for different clients (public entities, law firms, etc.), allowing users to access large amounts of information and reduce operational times.

BPS Process Automation, with Python REST APIs and data handling with pandas, allowing significant processing time reduction.

Development of a full-code multi-agent AI system in Python for queries on a BI system.

Implemented with LangGraph and LangChain for agent orchestration and conversational flow management.

Processes natural language questions and generates dynamic queries to business data. Information retrieval for context through a vector database (OpenSearch) and dynamic searches to PostgreSQL.

Integrated with MCP microservices connected to BigQuery, ensuring a modular and scalable architecture.

Development of a sentiment analysis system focused on price perception from product reviews.

Development of an advanced chatbot for Piotr Rusinek using the GPT-3.5 model, with the aim of generating new content based on his book "Before the Hammer Falls". The chatbot could create questions, generate accurate answers, and produce additional material such as slide content. To achieve this, the book was vectorized and stored in a vector database, allowing the model to select and provide relevant and contextual information.