Data Scientist

CloudWalk:

• CloudWalk is an AI first company building its own technology to bring justice to the broken payment system in Brazil. We are building what one would call a “self-driving bank”. 
• Some people say a company is a unicorn when it reaches a valuation of over 1 billion dollars. We are one of those companies. But we are also a much cooler type of unicorn: a freaking amazing beast that is incredibly rare, cool as hell and damn hard to catch.

The R&D team:

• We love data.
• We love living in a time when there’s access to an immense corpus of shared knowledge and incredible statistical and computational tools to extract meaningful information from the data.
• We love thinking in high dimensions.
• We like to explore before we exploit.
• We sprinkle sci-fi references in everything we do.

What the job entails:

• You will be part of an exploration team inside the AI department. We take a step back from the day-to-day urgencies and pursue ambitious projects with high impact potential.
• You'll have access to tons of data, but you'll also be cursed with tons of noise. Most humans are good and amazing, so every bad behavior is hidden behind a swarm of good behavior. To identify the baddies, you will need to look for clues as to what is going on, to look for patterns, to ask intelligent questions and look for answers, to deal with uncertainty, to find the story behind the data, to turn the data into information. You'll need to be a detective.
• You will create new ML models and techniques and expand on the ones we have in production.
• You will have the chance to experiment. Do you think that the hot new model from a newly published paper can be applied to some of our data? Let's try it! Do you think the atmospheric pressure, the migration pattern of birds or the number of capybaras living on the banks of the Pinheiros river are good predictors of credit card fraud? Let's investigate the data! Do you think there's a different perspective to look at a certain problem that could solve it better or complement our current approach? Let's put that to the test!
• You might have taken a bunch of courses and learned all about NNs, CNNs, RNNs, SVMs, CARTs, RFs, LDA, QDA, XGB, BERT, SD, GPT, and all that nice stuff. Those are all techniques to learn a mapping from X to Y. Here, you will also have to think long and hard about what X you'll be mapping to what Y.
• If you never allow yourself to fail, you will never allow yourself to take risks. The next big breakthrough will not come from someone who's playing it safe. Here, you'll be expected to experiment and you'll be allowed to fail, as long as you learn something. Over 99% of all species that ever existed on Earth are now extinct. But the remaining 1% would not have succeeded without them.
• Not going to lie: sometimes you will have to do things you might not particularly enjoy. There will be times when we'll be at war, when we'll be together in the trenches, fighting the enemy, shoulder to shoulder, doing whatever needs to be done.
• As a member of a fully remote and distributed team, you are expected to complete tasks autonomously, being highly collaborative and self-driven. We expect you to have the curiosity of a child with the responsibility of a grown-up.

Requirements:

• You must understand what machine learning is and what it’s not, and be able to identify and implement potential applications.
• You must be able to go through large amounts of data and identify useful patterns and relationships.
• You must be comfortable with programming. You’ll be doing most of your work in Python and SQL, but we believe that coding is mostly about problem solving and that programming skills are highly transferable between languages.
• You must be in your element reading and discussing research papers and coming up with potential applications to our problems.
• You must be able to communicate and debate in English. We’re proud to be a global team with members spread all over the world.  Portuguese is not required.

Recruiting process outline:

• Our selection method is simple but hard. If you pass, you are definitely smart.
• 1) Online technical assessment
• 2) Technical interview
• 3) Cultural interviews
• If you are not willing to do an online quiz, do not apply.