I worked as a I+D developer at the core team of Keepcon, a natural language processing (NLP) company that provides services such as moderation, topic discovery and sentiment analysis for social networks.

Data Scientist

I worked with a multidisciplinary team doing research about the brain activity of kids from low income neighbourhoods, trying to check the effectivity of different therapies and activities used to improve their cognitive abilities, I developed activities and games that kids used while the brain activity is recorded and synched with those programs. Some of the results of this research can be seen in this paper we published on november 2018 (https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0206983).

Software Developer

As a side project from my thesis I was proposed to work at the Thomas Jefferson University, as part of Shennan Weiss MD team, improving the processing time and code quality of some programs used in clinical medicine for patients with epilepsy. Those were used to detect HFO (High frequency oscillations) from EEG recordings, these oscillations are related to epilepsy so it’s used to locate the malfunctioning brain regions

My job in MercadoLibre consisted on improving and developing the backend of the billing software used by the site.

Software Engineer

I taught algebra, geometry and calculus courses.

High School Math Teacher

I work daily with Python, this is my favourite language. I've built hundreds of APIs, web scraping scripts and machine learning classifiers with it. I also contribute frequently to open source projects like FoxDot (https://github.com/Qirky/FoxDot/graphs/contributors)

I develop machine learning and data science products regularly.

You can check some works here:
Neuroscience experiments (https://github.com/mathigatti/Emotiv-Experiments)
Small data streaming website (https://github.com/mathigatti/WeatherSensorStreamer)
Testing different machine learning algorithms for spam classification (https://github.com/mathigatti/AA-TP1)
Clustering example (https://www.codementor.io/mathiasgatti/the-beauty-formula-identifying-interesting-patterns-automatically-based-on-aesthetic-metrics-basic-clustering-example-with-scikit-learn-xka5d6do8)

I scrap websites frequently in order to generate new datasets, extract useful information from websites or create programs that interact with the web.

You can check some related works here:
Midi to voice synthesis (https://github.com/mathigatti/midi2voice)
Normalized google distance implementation (https://www.codementor.io/mathiasgatti/python-implementation-of-normalized-google-distance-simple-web-scraping-example-vrwwu5w58)

As a software engineer, in MercadoLibre Billing team, I managed the databases that stored all the sales information of the number one latin america e-commerce site. I usually developed Java products to manage SQL databases through Hibernate, at the same time I frequently used complex queries to extract useful data, fix and detect bugs. I have also used Redis, MongoDB and other NoSQL databases in several projects.

Software Developer specialized in Data Science

Mathias Gatti

I worked on an interactive app which allows people to walk through museums and public parks and receive notifications with descriptions, audios and images when they are close to some relevant place. The first version was exhibited on march 2019 at the Jardin Botanico of Buenos Aires under the title of "Jardin Sonoro".

You can check and donwload the app in the attached link.

Proximity Marketing app using BLE (Bluetooth Low Energy) Beacons

Recently I became interested in singing synthesis so I programmed midi2voice, a small script to generate vocals from midi files.

After that I thought It would be useful for live performance, generating vocals in real time, so I added this feature to FoxDot, the famous live coding interface.

Personal details

Mathias G.

Summary

Technical skills

Work Experience

Personal Projects