I founded NaN to be a company that uses numbers and technology to solve problems and grow sustainable products.We are a small team, but we have already solved problems for big companies such as Coca-cola and Nestle through WMcCann, for the Brazilian civil society through one of the most relevant Brazilian think tanks, namely FGV, and for attorneys by growing products such as minutario.online.

At NaN Systems, we divide the consulting part of the job into two halves: effectiveness and efficiency. The former is about discovering the path while the latter is about the velocity of the travel.My job in this role is to understand the numbers and the company's vision of the future, design processes, and implement the tools to generate leverage.I work mainly with Python and Django on the backend. I've chosen these tools because of the fantastic community and because I can use Python to create a simple script and sophisticated Machine Learning stuff.In a world full of front-end frameworks, I prefer using HTML and CSS to bring user interfaces to life. I also use SVG and d3.js to build dashboards with valuable data visualizations and data exploration tools.It is worth mentioning two customers: - Metr, a Berlin-based company that gathers data from all sorts of sensors to monitor and optimize things like heating, power, and water consumption. There, we grow a multicultural team that delivers solutions constantly and efficiently, being able to handle every part of the stack ourselves, from the embedded code to the backend, from web apps to infrastructure optimization.- Aos Fatos, a fact-checking agency that plays a critical role in fighting fake news in Brazil. Our main achievements over there were: 1) to stabilize the main website, from a scenario of having daily downtime episodes to 100% of uptime; 2) to double the team's productivity; and 3) to release a new product to automate audio transcription for journalists (https://escriba.aosfatos.org/).

At Acordo Fechado, we design ways to solve more legal disputes and solve them faster.Our goal was to help law firms, courts, companies, and civil society, by eliminating waste of time and work while making the best agreement to the parties.My main contribution as the CTO of this ODR was converting months of waiting for nice-to-have features into a company that generates revenue by wisely applying the right amount of technology to each customer. Of course, I was not the only one responsible for that; the team was already incredible before my arrival, but I firmly believe that our skills match made the difference.

As a startup inside a big company, our small team used to design, build and maintain products that support Business Intelligence teams automating operational tasks, creating metrics, and using data visualization techniques to get business insights.We built dashboards for worldwide known brands and saved thousands of hours by automatically gathering social media data and generating spreadsheets and reports.

SegPub - http://segpub.pypln.org/Visualizing categorical and spatial-temporal data of security incidents from Rio de Janeiro to answer questions such as: • Where there were more injuries during the last week? • What has been the most frequent incident this month? • In which slice of the day there were more threats?Technologies used: Node.js and d3.js. Info Dengue - http://info.dengue.mat.br/This project aims to efficiently report and predict Dengue epidemics in Brazilian cities using time series analysis methodologies and data visualization techniques.I was responsible for creating data visualization tools and writing software to collect and parse meteorological, geological, and Twitter data to feed the model.We have Python, Pandas, Django, and R as primary tools.

Videolog is a Brazilian version of YouTube; it was born before the famous video portal but did not achieve as much success. I joined the company during its reboot, and my role in this process was both as a programmer and in spreading good programming practices.The scenario was full of challenges; there were no automated tests, we couldn't count on the Git repositories as trustful documentation, there were thousands of useless lines of code, severe security issues, critical parts of the systems were so coupled that testing them in a development environment became very difficult, and so on. In ten months of working hard with the team, we were able to:• Stop using Git as it was SVN and adopt a simplified version of Git Flow;• Develop a new REST API 100% covered by tests;• Deploy the first version of an in-house video encoder that was delayed for more than one year;• Increase the uptime of the primary product from 82% to 99%;• Deploy another MVP: http://puffer.tv;• Write setup scripts for almost all projects;• Automatize a bunch of operational tasks.We use Javascript (node.js) and PHP as the main languages, Ruby for a small Rails app and deployment scripts, MySQL and MongoDB as databases. All of that was hosted on Amazon's infrastructure.

I have played as Tutor of an experimental discipline that aimed to teach Python to law students.During the semester, the students were encouraged to think how to apply their recently learned skills to change the juridical reality. I am proud to say that some of them are developing their ideas and starting tech startups.

At the FGV Law School, we were usually required to answer a broad variety of questions. This way, I was able to apply and improve some Data Analysis techniques in a more pragmatic environment. At this point, it is worth to mention two research projects: • Supremo em números, a project that aimed to explain what happens at the Brazilian Supreme Court. I have applied Natural Languages Processing techniques to inspect judges' decisions and tell how they cite each other. Moreover, developing simple statistical analysis and administrating MySQL and Oracle databases were part of my routine. • Habeas corpus nos Tribunais Superiores, which analyzed how Habeas Corpus cases were handled at the Superior Court of Justice. To do this, I automatically collected data from the court website and hundreds of Excel spreadsheets filled by law students. But the most important part of my work was creating tools, such as D3js exploratory visualizations, that helped the lawyers ask new questions and get better results. For example, I made an optimized version of the Parallel Sets D3js implementation that could handle more dimensions and values per dimension. You can check it out here: http://israelst.com/parsets/tjsp.html

I was a member of the NAMD team, which stands for Data Analysis and Data Modeling Research Group. It is a multidisciplinary team comprising engineers, physicists, mathematicians, and other professionals who work collaboratively to develop innovative solutions involving natural language processing, data mining, and data visualization.My primary responsibility at NAMD was working on the Legal Meta Database project, also known as MBJ. This project was designed to collect and organize data on every lawsuit case in Brazil by crawling the State Courts Websites. Our initial goal was to download more than 90 million pages per month without disrupting the target website, and to parse these pages to extract complex key-value data.To achieve this, we employed distributed computing techniques using Python, lxml, and XPath for parsing. We used Bottle and D3js for data visualizations to track results and MongoDB to save each step of the data processing.

My work at BielSystems/Helabs was not just as a programmer; this was my first experience running an entire business with a friend. We had a typical Ruby on Rails stack back then, but the most crucial aspect of that time was learning the importance of automation at every level of a product and how to decide based on customer feedback. At that time, our products were: • http://cobregratis.com.br/ • http://www.webfax.me/

I was part of the team that accomplished one of the biggest website reformulations of ANS, the healthcare regulatory agency of Brazil. One of my main attributions was to connect the development and infrastructure teams, working to deploy, monitor and keep the whole portal.The biggest challenge was creating new features using two different frameworks, Lumis and Joomla, with a small team in an unstable political environment.

Software development with Ruby, Python, and Java.

My primary role was to maintain the company's main website and integrate it with the Franchise Management System.Used technologies: • Lumis, XSLT, XML, ASP • PHP • HTML, CSS, Javascript • Web Services • Test Driven Development

I performed tasks involving analysis and development in most of the company projects. Moreover, I was engaged in internal process refinements to improve team productivity and satisfaction, and the quality of deliverables.Technologies: • PHP, Drupal • Lumis, XSLT, XML • MySQL, Oracle • HTML, CSS, Javascript • SEOCustomers: • Silimed • Bradesco Seguros • Grupo CCAA • Instituto Euvaldo Lodi (IEL) • Confederação Nacional da Indústria (CNI) • Haztec • Norvartis

I have taught front-end development (HTML, CSS, and JavaScript) for three semesters. My emphasis was to raise awareness about web standards through a hands-on approach.

The task involves analyzing and developing a tool that integrates project management, billing, and automates the creation of documents for the company's core business.The key challenge here was navigating different areas in a medium-sized engineering company as a software analyst and developer. Having such an opportunity at the beginning of my career shaped the importance I give to agile principles, such as intensive and close interaction with the customer.

Software development for improvements in document management of the Production Department.This was my first professional experience. I was not working on a big project, no rocket science there, but, as a teenager in a military environment, I had to learn to combine creativity and discipline to deliver innovative solutions predictably.