Hi — I'm Ronny Toribio

Machine learning engineer — deep learning, computer vision, model deployment, and scalable inference systems.

Featured Projects

Coffeeless Programming Smart Search Service

A serverless Retrieval-Augmented Generation system that transforms educational articles into an intelligent semantic search experience. Built with AWS Lambda, Amazon Bedrock, and Pinecone, the platform indexes selected article sections, performs fast HNSW-based vector search, and uses an LLM to deliver accurate, context-aware answers in real time.

Python AWS Bedrock AWS DynamoDB AWS Lambda Pinecone AWS Cloudwatch Hierarchical Navigable Small Worlds(HNSW) approximate k nearest neighbors (kNN) cosine similarity Clean Architecture
View Project

Coffeeless Programming Quick Search Service

Implemented a search and autosuggestion service using TF-IDF, Levenshtein ratios for typo handling, and NLTK preprocessing to deliver accurate, context-aware article recommendations for the Coffeeless Programming educational platform.

Python Term Frequency-Inverse Document Frequency (TF-IDF) NLTK Levenshtein distance AWS Cloudwatch
View Project

Coffeeless Magic Tokens

The coffeeless magic tokens package provides widgets and react hooks for displaying streaming text. We use this package on the Coffeeless Programming Smart Search page for displaying decoded tokens from our RAG semantic search system using a glowing effect. Our glow widget has light and dark themes that match Coffeeless Programming's day and night themes. You can customize the colors using props. The widget also detects overflow and displays a button that allows you to expand and collapse the widget depending on your needs.

Typescript Javascript React TSX CSS
View Project