Below you’ll find a list of projects I built over the years. Some were personal projects, others were developed for one of my companies. The list is not complete, but it gives a good idea about my interests.
Despegar.com Chatbot (2017): Despegar.com is the largest online travel agency in South America. They hired me to create a chatbot that allows their users to receive travel information and search for flights directly on Facebook Messenger.
I coded everything in this project, including the Natural Language Processing algorithms (i.e., the software that tries to understand the text messages of the users). It was a lot of work, but now it is a blast to watch my bot chatting with humans. They sure try to give him a hard time!
BolsaCapital.com (2015): Throughout 2015 MasterCard is hosting a series of programming competitions around the world called “Masters of Code.” Competing teams need to develop a product that uses the MasterCard APIs to solve a problem in the financial sector. My team won the São Paulo edition with a project called “Bolsa Capital,” and now we’ll be flying to California to represent Brazil at the grand finale.
Small business owners in Brazil need to wait an average of 30 days to receive the payments for goods and services they sell via credit cards. They can anticipate those receivables, but only with the credit card acquirer, which creates a kind of a monopoly. The goal of our platform is to allow those business owners to negotiate directly with several banks at the same time, online and with a couple of clicks. We believe such a platform will bring more transparency and efficiency to this huge market segment, reducing the interest rates the small business owners need to pay. You can read more about the competition itself here.
Good Driver App (2015): Early in 2015, during the Campus Party event in São Paulo, Ford organized a Hackathon where competitors would have 24 hours to develop a mobile app that interacted with its Sync AppLink technology (a platform that provides data about the car and allows users to control their smartphones via voice commands). The prize was a brand new car, and the apps would be judged according to three criteria: innovation, technology and business model.
I developed an app called Good Driver (from Portuguese “Bom Motorista”). The goal of the app is to make the process of evaluating insurance risks more efficient, thus allowing insurance companies to charge a more fair price to good drivers. The app basically collects a wide range of data about the driver (e.g., average speed, acceleration pattern, most used streets, most frequented locations), creating a driver-profile. That profile is then used by the insurance companies to calculate prices more efficiently. As you can guess by the picture above, I ended up winning the competition, and I am currently working on a beta version of the app to test the market viability. If you want to know more about the Hackathon itself, check out this detailed post I wrote about it.
Suggestion Box App (2015): The idea for this app came to me when I was paying the bill at a restaurant and I noticed a plastic “Suggestion Box.” I had a couple of suggestions for the restaurant, but I sure didn’t want to waste my time grabbing a pen and writing down my feedback. Other people surely felt the same, cause the box was empty. I figured there should be a more efficient way to solve this problem.
The Suggestion Box app (from Portuguese “Caixa de Sugestão”) allows customers to review their favorite stores, and to send criticism and suggestions, directly from their smartphones. In order to encourage the reviews and suggestions, whenever a customer sends a review, he receives a discount coupon for his next purchase at that same store. Reviews and suggestions are private, and store owners get to see all the reviews, with averages and monthly charts, on a web interface. I am currently testing the app on selected stores in Brazil.
AppRush.com (2014): The idea was to create a marketplace connecting mobile developers who need feedback on their new mobile apps, with users who are willing to provide it. Developers would pay a fixed amount per feedback, and the users would get part of that money for downloading, testing the app and filling out a questionnaire.
Some of my test clients asked if the users could post reviews directly to the app store, and I figured that if I wanted the marketplace to grow I would need to offer that service. The problem is that offering app store reviews, even if you guarantee honest and real reviews, is a gray area. First of all it could create problems with Apple and Google. Second, it would get my service mixed with shady ones that offer hundreds of positive reviews for $5 (i.e., fake reviews). I was not comfortable with either of those situations, so I decided to shut down and move on.
MobiCard App (2013): Have you ever lost one of those paper loyalty cards? I am sure you have. Thinking about that problem, in 2013 we (my brother and I) decided to launch a loyalty program that was 100% mobile based, called MobiCard. Customers would install an app, make a purchase at their favorite store, and the loyalty points would automatically get added to their accounts.
Merchants had complete control over the rewards offered, how the points were earned, and they could even notify customers about special deals and promotions. We launched the app with a customized version for a couple of test clients, and the feedback was positive. Right after that we run into some disagreements regarding the direction we should take the project, so we decided to drop it.
Carriers Test App (2012): During the Intercon 2012 conference I participated on another Hackathon. The goal was to create a mobile app that would have a positive social impact. Along with my brother (who is also a programmer) I created an app called Carriers Test (from Portuguese “Test das Operadoras”). The idea was to measure the quality and speed of the mobile Internet connect provided by all the carriers, and then make this data public, so that customers would be able to make the best choice according to their region, and that carriers would be encouraged to provide a decent service, or else they would be called out on it.
We got the second prize at the Hackathon, and afterward we tried to bring the app to the market. Given that the data we would release could have an economic impact upon the carriers, we wanted to make sure we were safe under a legal point of view. We therefore looked for ways to get a certification from local regulatory bodies, but the process became too expensive, and we decided to drop the project.
TwitFever.com (2010): In 2010 Twitter was the place to be, and Internet marketers and online professionals became obsessed with having powerful and popular accounts. Identifying this opportunity I launched TwitFever.com, which offered a set of tools to manage one’s Twitter account.
One of the tools allowed the client to specify keywords, and the application would automatically follow users who had used that keyword in recent tweets. After a while Twitter prohibited such automatic following of users. Given that this was one of the most used features of the service, I decided to shut it down and sell the domain name.
CopyTrust.com (2009): Back in 2009 duplicate content was a serious issue for webmasters and website owners. Google could penalize a certain site if its content was appearing on other sites as well, as it was not always easy to identify who was the copyright holder. A consequence of such penalty could be the loss of search engine traffic. In order to combat this problem I launched CopyTrust.com, a service that would automatically monitor the web for sites copying your content, and not only that, it would also automatically send DMCA take-down notices to the offenders. It turned out that website owners were not willing to pay much for such a service, so I had to close shop.
OnlineProfits.com (2008): Online Profits is the name of my first company, and I used the website to offer a 6-month course about Internet Marketing. I recruited top experts from a wide range of backgrounds to create the different modules of the course, and the result was a pretty comprehensive and advanced program.
The coursed grossed over $300,000 in revenues during its first two years. After that I got involved with other projects, so I stopped working on the course directly. The program is still available on the market, but at a reduced price, because students don’t get direct support from the experts anymore.
DailyWritingTips.com (2007): My goal with this website was to help people who want to write clear, correct English. Every day we publish a new article, with topics including style, spelling, grammar and punctuation. Today the site receives over 2 million visitors per month, and has 100,000 email subscribers.
The site was feature on several mainstream publications, including the Wall Street Journal, The Chicago Tribune, The Huffington Post and Writer’s Digest. In addition, our articles are mentioned or recommended by over 30 US universities.
IAS Simulator (2012): The IAS was the first electronic computer, built by John von Neumann at Princiton’s Instutite for Advanced Study. As a project for a CS class I was taking I built a simulator of the IAS computer entirely in ARM Assembly.
It was an awesome learning experience, as it allowed me to understand important principles behind any computer architecture, like the benefits of storing programs alongside data, the fetch-and-execute cycle, the basic structure of a microprocessor, and so on. You can see the code of the IAS simulator here.
4-bit Adder (2011): After reading Code: The Hidden Language of Computer Hardware and Software, I decided that I wanted to build my own bit adder, using only wires and relays. I started with a simple 4-bit adder, which you can see below. The next one will be an 8-bit one that can perform both addition and subtraction, but I’ll be using transistors instead of relays for this one, else it would take years to complete!
Java Web Robot (2010): Quite frequently over the years I came across a website or web application where I needed to perform the same task over and over again, but site in question didn’t offer such bulk action option. For instance, whenever you send a marketing email with SendGrid, it will produce a list of email addresses that bounced. The problem is that there is no way to mass delete those bounced emails from your list. You need to do one email address at a time.
To solve the problem, back in 2010 I wrote a little Java script that creates a web robot, automating such tasks. I just need to configure where and how the robot is supposed to click and type, where it should get its data from (e.g., local files or other websites) and it does the rest by itself, simulating the behavior of a human user. It’s quite amusing to the see robot in action, and it saved me hours of time by taking care of those repetitive tasks over the years.