Machine Learning Engineer2018 - 2020Migacore
Technologies: Python, Jupyter, Pandas, Scikit-learn, SpaCy, XGBoost, Spark, Docker, Kubernetes, Scrapy, React
- Built the infrastructure for collecting structured and unstructured data from the web and social media.
- Used NLP techniques to extract events from unstructured text and understand their importance.
- Built data processing pipelines using Python and Spark.
- Explored, visualized, and cleaned up airline data using Pandas and Matplotlib.
- Developed machine learning models for assessing event importance, and for predicting travel demand and price sensitivity.
- Built a web app for showing important events and their impact on travel; used React and Redux.
- Deployed data pipelines and web apps on GCloud and AWS, using Docker and Kubernetes.
- Mentored other developers—both on the full stack and the machine learning side.
Full-stack Developer2014 - 2017Freelance
- Designed and implemented a small CMS using Express.js.
- Developed responsive websites using the aforementioned CMS.
- Extracted and processed data from third-party APIs (Facebook API, Google Maps API, Eventbrite, Meetup, etc.).
- Scraped websites and processed data about events or places.
- Developed mobile apps using Ionic, Cordova, and Parse.
- Deployed and maintained apps and websites on Parse, AWS, and Heroku.
- Designed, implemented, and interpreted complex analytics using Google Analytics.
- Designed and implemented, in collaboration with an on-site team, several components in a complex web ERP framework built with Angular and .NET.
- Researched and proposed architectural changes for the aforementioned framework.
Front-end Developer2014 - 2014Recognos
- Chose the technology stack and created the architecture for the single page application developed in AngularJS.
- Designed and implemented a tool for semantically annotating an HTML article.
- Refactored and extended a legacy KnockoutJS application.
Co-founder and Developer2012 - 2013Orderick
- Identified financing opportunities and drafted a business plan.
- Pivoted several times to identify the most suitable MVP.
- Researched and employed linked data standards and libraries.
- Extended BackboneJS to build a framework for dynamically handling semantic data.
- Used third party APIs to gather data and integrate content.