Lead Developer2018 - PRESENTPinTecnologia
Technologies: Django, React, Ethereum Blockchain
- Built a system to process administrative documents over the Ethereum blockchain.
- Worked on the front-end and built the UI using React.
- Designed and developed the back-end using Django and the Django REST framework.
Product Owner2017 - PRESENTBMD Software
- Managed the PACScenter product which is an all-in-one medical imaging platform for patient studies (storage, visualization, and sharing)—enabling simple and efficient workflows.
- Developed core features for both the back-end and front-end.
- Defined the strategy for new version releases.
Kaggle Competitions Expert2018 - 2018TalkingData AdTracking Fraud Detection
Technologies: Apache Spark, MLlib, Gradient Boosting Machines, Feature Engineering
- Worked with a large dataset using big data tools like Apache Spark.
- Built an algorithm that predicts whether a user will download an app after clicking a mobile app ad which helps to combat click fraud.
Kaggle Competitions Expert2018 - 2018Toxic Comment Classification
Technologies: Deep Learning, LSTM, GloVe, Embeddings, LSTM, GRU, Ensembles, Naive Bayes, SVM
- Studied negative online behaviors e.g., toxic comments.
- Built a multi-label model that’s capable of detecting different types of toxicity like threats, obscenity, insults, and identity-based hate.
- Used a labeled dataset of comments from Wikipedia’s talk page edits.
Kaggle Competitions Expert2018 - 2018Recruit Restaurant Visitor Forecasting
Technologies: LSTM, ARIMA, Time Series, Gradient Boosting Machine
- Predicted how many customers to expect in each day in a restaurant to effectively purchase ingredients and schedule staff members.
- Developed a prediction model for this task; this was not easy to make because many unpredictable factors can affect restaurant attendance like the weather and local competition. It's even harder for newer restaurants with little historical data.
- Worked with heterogeneous datasets.
Kaggle Competitions Expert2017 - 2017Sberbank Russian Housing Market
Technologies: Gradient Boosting Machines, Ridge Regressor, Feature Engineering, K-fold Cross Validation
- Created a prediction model capable of making predictions about realty prices so that renters, developers, and lenders are more confident when they sign a lease or purchase a building.
- Developed algorithms which use a broad spectrum of features to predict realty prices, using a rich dataset that includes housing data and macroeconomic patterns.
Kaggle Competitions Expert2017 - 2017Two Sigma Financial Modeling
Technologies: Reinforcement Learning, TensorFlow, Extra Trees Regressor, Ridge Regressor
- Applied technology and systematic strategies to financial trading in order to forecast economic outcomes that can never be entirely predictable,.
- Back-tested to validate regression models that predict financial time series.
Kaggle Competitions Expert2016 - 2016Santander Customer Satisfaction
Technologies: Random Forest, Decision Trees, Ensembles, Neural Networks, PCA
- Created a model that identifies dissatisfied customers.
- Worked with hundreds of anonymized features to predict if a customer is satisfied or dissatisfied with their banking experience.
Kaggle Competitions Expert2014 - 2014Loan Default Prediction
Technologies: Random Forests, Gradient Boost Machines
- Determined whether a loan will default, as well as the loss incurred if it does default.
- Developed methods unlike traditional finance-based approaches to this problem, where one distinguishes between good or bad counter parties in a binary way, we sought to anticipate and incorporate both the default and the severity of the losses that result.
- Built, as a team, a bridge between traditional banking, where we are looking at reducing the consumption of economic capital, to an asset-management perspective, where we minimized the risk to the financial investor.
Kaggle Competitions Expert2013 - 2013Job Salary Prediction
Technologies: Random Forests Regressor, Ridge Regressor, Bag of Words, TFID, Dimensionality Reduction
- Built a prediction engine for the salary of any UK job advertisement so they can make huge improvements in the experience of users searching for jobs, and help employers and job seekers figure out the market worth of different positions.
- Worked with a large dataset (hundreds of thousands of records) which was mostly unstructured text with few structured data fields. These were in a number of different formats because of the hundreds of different sources of records.