Using Deep Learning and Reinforcement Learning methods, I am working on building anAuto Machine Learning system that automatically searches for models and derives end-to-endpipelines that read, pre-process the data, and train the model.

Peruvian startup (https://doctor-cv.com/) which evaluates and proposes improvements inyour CV using NLP techniques. It is like having a virtual recruiter expert who will tell whatto change in the resume to increase the chances of being called for the job interview.

I worked on extracting symbolic knowledge about objects and their characteristics from unstructured text (relying on natural language processing, machine learning and machine reading techniques), as well as available ontologies and knowledge on the Semantic Web.

I researched some information retrieval techniques with machine learning methods and word embedding representations to create hierarchies of terms in collections of patent about different topics. In this project, I used Spark, HBase and Python.

This project was a partnership between ICMC-USP and SAMSUNG Eletrônica da Amazônia LTDA, its objective was to advance the state of the art for semantic processing of texts/documents written in Brazilian Portuguese.

I served as TA to professor Dr. Thiago Pardo for the course "Professional Information in Computer Science". I was responsible for evaluating and grading examinations or assignments. In collaboration with another TA, we supported 50 students during discussion sections and office hours.

I implemented a crawler for Twitter in Python. This crawler collected public comments about music, books and films. I also developed a module to normalize these comments (spell checker, etc.). With this data and other resources the database was increased.

I researched and developed some automatic techniques of opinion mining and sentiment analysis for Spanish language. In addition, I implemented a crawler for Facebook and Twitter in Python. This crawler collected all public comments written on these social networks. Lindexa was a winner startup of the first edition of Wayra Peru.

In general, we researched techniques to generate extractive summaries of many texts. These summaries selected the most important data in terms of coverage and representativeness. I also researched and implemented text classification and clustering algorithms.

We developed the Graphical User Interfaces of the Credits and Collections module, it included the elaboration of the navigability documents and prototypes. I also developed the executable prototype.