Data Scientist
2015 - 2018Major Software Company (Redmond, USA)- Identified a reseller’s product bundles had a high level of customer churn; recommended certain modifications for the reseller offerings which resulted in a 5% reduction in churn rate.
- Built propensity models that were the basis of marketing campaigns and which generated millions of dollars of new recurrent revenue.
- Experimented with my user segmentation model and feature recommender which resulted in an increase of the treatment group's user engagement.
- Designed and developed a tool to quickly create, test, and operationalize end-to-end churn prediction models for the SMB Office 365 subscribers.
- Developed a patent-pending clustering method of Excel commands based on co-occurrence in the command sequence. The algorithm allows product teams to categorize features and customers.
- Wrote an efficient algorithm to segment users according to their level of proficiency in Office. The pipeline process terabytes of clickstream data. Also demonstrated that increased proficiency is a driver of subscriptions renewals.
- Constructed an in-application recommender for Office client; the model is based on a user’s level of proficiency, used and unused commands, and similarity between users and commands.
- Identified and quantified biases of a Network Promoter Score survey for Outlook compared with other office apps; used machine learning techniques which identified drivers of NPS and quantified the prevalence across different populations. Also proposed a methodology to measure and correct bias.
- Participated in multiple hackathons organized by the machine learning community at Microsoft. More information can be found in my portfolio.
Technologies: Python, R, C#, Machine LearningTeaching Asistant
2013 - 2016Harvard University- Acted as a teaching assistant in 2013, 2014 the data science course; grading coursework and mentoring students.
- Assisted and taught in 2017 the machine learning course; grading coursework and mentoring students.
- Developed Python notebooks with didactic examples and sample code so students can easily absorb topics such as linear regression, bias vs variance concept, cross-validation, and Random Forests.
- Mentored students on their final projects and helped them to be successful by evaluating project proposals, scheduling follow-up meetings, and grading project presentations and final reports.
- Constructed a guide to install TensorFlow on the Windows platform using Docker containers and Hyper-V.
Technologies: Jupyter Notebooks, Python, SklearnSenior SharePoint Support Escalation Engineer
2009 - 2015Major Software Company (Redmond, USA)- Dealt with the most critical and complex customer’s issues involving numerous Microsoft technologies.
- Analyzed large quantities of product and weblogs and applied visualization techniques to communicate and influence the customer to perform actions toward case resolution (implementing a number of data mining techniques).
- Significantly improved customer satisfaction among Latin American enterprise customer resulting in renewal and scale-up of premier contracts.
- Engaged the product group to resolve high-business impact product bugs that I had identified during case troubleshooting phase. The resolution of them, during the scale-up, improved the product localization for Spanish and Portuguese speaking customers.
- Constructed a team of engineers to expand the language support of SharePoint Online and OneDrive and as a result, customer satisfaction among Spanish and Portuguese speaking customers improved significally.
Technologies: SharePoint, Windows Server, C#, WinDbgSenior Consultant
2006 - 2009Major Software Company (Redmond, USA)- Developed that complete business intelligence solution for an important government organization; the solution consisted of three cubes and dashboards supported by SharePoint. Worked with a customer’s analytics organization to understand their requirements, and finally design and implement the solution.
- Implemented CETS, an innovative collaborative law enforcement solution based on SharePoint and .NET.
- Worked on the CETS initiative from incubation to final deployment; CETS is a global Microsoft initiative implemented across the world supporting different languages and cultures.
- Modernized the home banking solution of an important bank and refactored the logging system from synchronous to asynchronous.
- Deployed the project server and project portfolio server in a major mining company.
- Collaborated with members from the major mining company define the right metrics to monitor and measure ROI of project and project porfolio.
- Designed and coded an InfoPath and SharePoint solution for a large goverment institution.
Technologies: C#, Agile Methodologies, SharePoint, SQL Server, Project Server, Analysis Server, Visual StudioCo-founder | Software Architect
1997 - 2005STI S.A. (Sold to Soluziona, currently Indra)- Designed and developed a COM+ component in an ATL 2.0 library, C++, and a Windows TCP/IP library. This asynchronous multithread connector opens two sockets to send and receive transactions and the load spikes were managed by queueing client requests.
- Designed and built a COM+ component in ATL 2.0 library C++ to allow a connection to a proprietary mainframe system. The COM+ Window Service managed multiple instances of the client APIs DLL individually in a separate process with a shared state through a SQL database.
- Created domain-specific InterDev design components; specifically, design time controls developed in Visual 6.0 to read transaction definition metadata and generate code to invoke, parse, and display transaction results.
- Envisioned, designed, and coded a content management platform in Visual Basic 6.0 and ATL 2.0 library C++. This simple template language enabled the graphic designer to create simple and attractive pages while the page content is stored in a SQL database.
- Developed a high-performance solution to process social security contributions directly from an employer's payroll. Employers can upload, validate, and process payment by transferring money from their bank accounts.
Technologies: ASP, COM+, Windows
Claudio Aldana
Software Developer in Seattle, WA, United States
Member since June 15, 2018
Claudio is a data scientist who focuses on business outcomes along with having a solid engineering background. He's applied the principals of data science to optimize customer satisfaction, increase product personalization and engagement, decrease customer churn, and improve customer segmentation. He has worked on the entire project lifecycle from scoping to deployment in production and approaches problems in a practical and agile manner.
Claudio is now available for hire
Portfolio
- Major Software Company (Redmond, USA)Python, R, C#, Machine Learning
- Harvard UniversityJupyter Notebooks, Python, Sklearn
- Major Software Company (Redmond, USA)SharePoint, Windows Server, C#, WinDbg
Availability
Part-time
Preferred Environment
Git, Python, R, Jupyter Notebook, Spyder
The most amazing...
...thing I’ve done was to improve the ROC-AUC of a customer churn classifier using LighGBM—it went from .77 to .85.
Employment
Experience
- Customer Churn Training and Prediction Pipeline (Development)
I designed and developed a tool to quickly create, test, and operationalize end-to-end churn prediction models.
- Participated in Multiple Machine Learning Hackathons (Other amazing things)
I participated in multiple hackathons organized by the machine learning community at Microsoft.
· A binary classifier to predict probabilities of upgrades for our Xbox 360 customers.
· A binary classifier to predict user attributes (gender, income, so on) based on BING search keywords.
· A text classifier to predict routing of new Office365 support cases based on an unedited customer description.
· A binary classifier to predict the expected probability of display add click. - Customer Segmentation for Office (Development)
I developed an algorithm to segment million of users according to their proficiency score using Office. I also demonstrated that an increased in proficiency score is positively correlated with subscription renewals.
- Command Recommendation System for Office (Development)
Using collaborative filtering techniques, I developed a command recommendation system for Office.
- NPS Sample Bias Analysis (Development)
I identified and quantified biases of Network Promoter Score surveys for Outlook compared with other office apps. Using machine learning techniques, I identified the drivers of NPS and quantified their prevalence for different apps, which allowed me to estimate score differences due to sample bias. I also proposed a methodology to measure and correct the bias.
- SharePoint and Power BI Solution for the Government (Development)
I developed a complete business intelligence solution for a major government organization. The solution consisted of three cubes and dashboards supported by SharePoint. I worked with the customer’s analytics organization to understand their requirements and design and implement the solution.
Skills
Languages
SQL, Python 3, R, UML, C++, C, Java, C#Libraries/APIs
SciPy, Sklearn, NumPy, Flask-RESTful, Caret, Ggplot2, Spark ML, PySparkOther
Software DevelopmentFrameworks
Ruby on Rails 3, Flask, SparkTools
GitHub, Rational Rose, Microsoft PowerPoint, Excel 2016, Microsoft Power BIParadigms
AgilePlatforms
Visual Studio 2016, RStudio, Jupyter NotebookStorage
SQL Server Analysis Services (SSAS), SQL Server 2008, MySQL
Education
- Master of Arts degree in Information Technology2011 - 2016Harvard University - Cambridge, MA, USA
- Master's degree in Business2005 - 2006ESE Business School - Santiago, Chile
- Bachelor's degree in Engineering1989 - 1996Pontificia Universidad Católica de Chile - Santiago, Chile
- Master's degree in Engineering1089 - 1996Pontificia Universidad Católica de Chile - Santiago, Chile
Certifications
- Deep LearningJANUARY 2017 - PRESENTDeeplearning.ai via Coursera
- Data ScienceFEBRUARY 2015 - PRESENTHarvard University
