Sao Jose dos Campos,
Brazil
(
GMT )
About
With a great interest in Data Science and Market analysis, throughout the years I've learned several tools to harness insights through data and had the opportunity to contribute with several distinct projects. Worked developing business intelligence projects and automation, supply/demand optimization, database creation and data pipeline design. Lately I've been working as Data Science Project Manager for Shooju. Shooju is a customized ETL/Machine Learning/Data Vizualiation PaaS with clients in the US and Brazil.
With a great interest in Data Science and Market analysis, throughout the years I've learned several tools to harness insights through data and had the opportunity to contribute with several distinct projects. Worked developing business intelligence projects and automation, supply/demand optimization, database creation and data pipeline design. Lately I've been working as Data Science Project Manager for Shooju. Shooju is a customized ETL/Machine Learning/Data Vizualiation PaaS with clients in the US and Brazil.
Experience
- Periodic calls with clients to develop project deliverables. - Scrum Master for Python ETL team, fitting client's demands with team tasks. - Coordinated a All-In-One Machine Learning product for clients. The project is now on expansion and consists of a non technical client friendly UI for multivariate time series forecasting. The project uses different models to evaluate the best one applicable, allowing a good market prediction for strongly correlated variables. - Coordinated imports and exports importer inferences project based on limited freight income data.
- Project coordination for the Dynamic Pricing Multi-Armed-Bandit algorithm. - Contributed to the development of bayesian AB tests to statistically test new website marketplace features, in order to raise average revenue per user. - Contributed to the marketplace's ranking algorithm core new features. - Developed a fraud detection statistical model based on conversion rates from the marketplaces partners. The model would analyze the prices and conversion rates correlation over a time series to identify potential suspicious foul plays. - Acted as Scrum Master for Business Intelligence team.
- Market analysis to prospect and plan offer and demand indicators, as well as wordwide detailed follow ups on the PA6 and PA66 feedstocks. - Implementation of a programming routine for a big client's contract price modelling using parameter variation and least squares error. The simulation allowed good prices forecasting using Adipic Acid upstream prices provided by IHS Markitt. - Monthly market share KPI's presentation, supporting commercial intelligence decisions