Projects ranging from model building to data evaluations and R&D at an identity risk management company• Researching potential of graph-based feature generation for future models, testing various graph embedding algorithms for performance improvement and production feasibility• Co-authored a paper with colleague on new approach to graph embeddings for ECIR 2020 conference• Worked on an updated machine learning credit model for Auto Industry, achieving KS improvement of 9-53% (depending on the client) over previous model• Development and presentation of new metric for measuring Disparate Impact, allowing clients to speed adoption of new models. Presented to clients at company’s annual conference in NY• Led project to evaluate coverage & value of internal & 3rd party data sources, created Python/Spark package to automate evaluations, and presented to internal stakeholders to evaluate ROI of data vendors

I assist the instructional team in the Data Science Immersive program, helping students with a wide curriculum ranging from statistics to NLP to unsupervised learning.

Galvanize’s data science immersive program is a 3-month, 700+ hour program designed to turn analytical and technically-savvy professionals into data scientists. Topics covered include Python for machine learning, regression, classification, NLP, recommendation engines and Spark on AWS. Students learn industry-ready skills through hands-on experience in case studies and projects.For my capstone project, I decided to build a model to predict whether a bill will pass or not in the California State Legislature. The goal of this project is to figure out which bills are mostly likely to pass so that citizens, NGOs and other interested parties can prioritize which bills they should focus on. I also wanted to see if there were patterns in the data that could reveal insights about the legislative process.I used a MySQL database made available by the Legislature that contains three types of data about bills: metadata (such as when a bill was introduced), author data (such as the party of the author), and the text of the bill. I found that most of the variation came down to three things. The first two were more obvious: party and taxes. The most interesting patterns emerged from examining latent topics extracted from bill text through a TFIDF to NMF dimensionality reduction pipeline. My final model was implemented in Python, using a Random Forest ensemble with a max depth of 8 so that it could take into account interactions between the 100 latent topics I extracted and party and other metadata features. The baseline I compared my model against was the 44% overall pass rate of bills, which yields an F1 score of .44 and an accuracy of .51. My Random Forest model was able to achieve an F1 score of .71 and an accuracy of .62.

Responsible for various social media data collection and analysis projects at a digital PR firm• Led several data-driven prestige projects to bring positive PR to firm resulting in 10+ major press mentions• Performed semantic and content analysis of social media posts using R to generate insights for clients on how to generate more engagement• Built crawlers in Python using Facebook, Twitter & Instagram API’s to collect data and managed MySQL databases with 3 million + data points

Social media data analyst role in entertainment vertical at a marketing analytics company• Created regular, detailed social listening reports for clients with data-driven insights and recommendations • Introduced new methods of analysis to generate more specific recommendations • Automated data workflow with simple web macros, reducing data wrangling time by 20%

• Adtech startup ViralGains is Mass Challenge $100K winner & 500 Startups participant in Boston, MA• Prospect leads at beginning of sales cycle, including cold calls and emails to potential clients• Assess BD processes and provide suggestions for optimizing sales strategies • Conduct competitor research and analyze new market opportunities, including a detailed report on opportunities in the growing Hispanic advertising sector

• Key member of outreach campaign to publicize the Greenvolved platform (www.Greenvolved.com), which achieved a social reach of 746,000+ people• Recruited 15+ university departments and organizations, as well as online bloggers and influencers to sign on as supporters of Greenvolved

• Support the Technology Products group in the School Achievement Services division at Pearson in Boston, MA• Perform extensive market and competitor research to advise development and marketing of new training• Investigate new EdTech products and how to integrate them into classrooms and Pearson trainings • Help guide discussion and development of new online platforms for e-learning

• Write content for blogs and social media to help improve individuals' and business' online reputations• Analyze clients' online and social media presence • Translate and adapt content between French and English

• Launched test prep branch in Laguna Niguel, CA as part of Revolution Prep's Ivy Insider's program• Gave courses and tutored high school students to help them improve their SAT scores• Brought in over $13,000 in revenue during the summer• Developed familiarity with P&L statements and CRM systems• Utilized numerous sales & marketing techniques, including personalized mailings, cold calling, blogging, e-mail and social network marketing

• Member of team working to revitalize company's marketing efforts• Perform interviews and market research to determine Bixby's real value proposition and brand• Develop inbound marketing campaign to generate more qualified leads

• Manage and maintain the website for the Peace and Justice Studies program at PJS.tufts.edu• Create a weekly newsletter sent to students and staff• Publicize the program and its activities, especially through social media• Assist in office work, filing, and reorganizing projects

The Resident Counselor provides supervision and guidance to the International Student Cultural Exchange (ISCE) students during the residential program. Specific responsibilities are as follows:Leading daily field trips and hall meetings.Assisting with activity planning and implementation.Helping with set up and breakdown of conference facilities and summer offices.Attending staff meetings as necessary.Leading student orientation and ensuring successful adaptation of students to life at Tufts.Other duties as assigned.

• Worked as counselor at summer day-camp (Lake Mission Viejo, CA) and after-school program at bilingual Las Palmas Elementary School (San Clemente, CA)• Completed extensive job training in dealing with customers, maintaining professionalism, group leadership, and conflict resolution• Individually held responsibility for a group of 20 fourth and fifth grade students• Lead activities, games, arts and crafts, tutored students with homework, coordinated group games and activities