Casai is Mexico's premier hospitality startup, aiming to revolutionize hospitality in Latin America by providing hotel quality along with AirBNB locality and accessibility. As part of the Data Science team, I have had the opportunity to extract value from data in several areas of the company, including an automated pricing optimization tool, controlled experiments to verify business hypotheses and inventory automation.
A suite of tools for the film social media platform Letterboxd, including a random movie picker, film taste compatibility test and a fine-tuned recommendation system based on latent matrix factorization. Use of several intersecting technologies, including the Letterboxd API, Streamlit, FastAPI, Docker, and AWS ECS make this a great case study as a complete end-to-end ML app. Also, users love it!
As part of a four-person consulting team of engineers from VTEX and outside data scientists, I built a universal, store-agnostic recommender system that is able to create recommendations for any one of VTEX's 2500+ e-commerce stores. Our method uses adapted NLP techniques to generate recommendations through a neural network and outperformed the baseline we set by 350-500%.
As Distrito Dataminer's Data Lead, I created and managed all of the infrastructure and technology that allowed us to become the most comprehensive and in-depth resource on startup data in Brazil. I also explored the data to help make Dataminer a profitable area within Distrito through client work and generated novel analyses that made it so it was cited in the national media over 160 times in 2019.