Developed full-stack features for an AI-powered platform utilizing Retrieval-Augmented Generation (RAG) to revolutionize content creation for presentations.

Implemented new features by building and enhancing the FastAPI backend, integrating OpenAI models through LlamaIndex, managing embeddings using PGVector (PostgreSQL), and fine-tuning prompts to optimize the relevance and clarity of LLM outputs. Additionally, developed the Next.js frontend using TypeScript and Tailwind CSS to create a seamless and user-friendly experience.

Boosted presentation generation speed by nearly 40% by parallelizing LLM calls using concurrent threading and optimizing the selection of LLM models to enhance performance while reducing costs.

Wrote end-to-end tests with Checkly and Playwright and integrated them into the CI/CD pipeline using GitHub Actions.

Developed and deployed a web application for geospatial data visualization using TypeScript, D3.js, Firebase Hosting, and Google Cloud Storage, enhancing data accessibility and introducing a new glyph-based methodology for urban population mobility analysis. This was done in the context of the government-funded project Fondecyt 1211490 ”The Geography of Segregation Based on Mobility Patterns Within the City: An Approach from Big Data and Complex Networks Analysis”.

Optimized cloud storage utilization by 60% by transforming raw data into Protobuf binary format, reducing storage costs and improving data retrieval speed.

Optimized SQL queries for mobile phone record analysis, boosting performance and accelerating data-driven insights into urban population movement patterns.

Full-stack Development with a focus on Microservices.

Implemented new features and resolved bugs in Ruby on Rails backend microservices and the React frontend, enhancing user interactions and web application functionality within the Scrum agile framework