- Increased average detection of grants by 300/week by adding processing of unstructured, social media data into existing ETL pipeline and accounted for duplicate opportunities found through our news sources- Assessed the accuracy and quality of a new data source before incorporating it into the training data for a news classification model, leading to a reduction of fake news in our philanthropy news feed by 27%- Evaluated news classification model on conceptual soundness, assumptions, and limitations, and presented insights and recommendations with data science team to the board of directors- Extracted features from text to use for classification, entity extraction, and record linkage- Assisted data scientists through full stack development: model deployment, retraining, maintenance, versioning, experimenting, and monitoring. Further interpreted and summarized results to stakeholders

- Automated data extraction, cleaning, labelling to improve fuzzy string matching of grant records by 12%- Partnered with other teams to create business requirements and communicate insights with visualizations- Designed experiments to challenge ML model limitations and documentation, and analyze user behavior- Conducted literature reviews and qualitative research- Tracked and reported website analytics; prepared summary data tables for analysis- Curated a digital library of case studies, evaluations, and reports and uploading files into an online database

- Mapped literature, policy, and key actors engaging violent extremism- Used a behavioral analytics platform to create psychographic profiles of alt-right subgroups- Tracked online radicalization using TF-IDF and sentiment analysis on Twitter data- Establishing thresholds for multi-edge graphs representing socio-political associations between subgroups

- Educated researchers on core computational skills: libraries, computational problem solving, best practices design patterns, test-driven development, and version control- Led office hours on code review, dataset analysis and interpretation, and establishing business needs

- Produced architectural renderings and diagrams to communicate designs and successfully acquired funding from pitch to Walter P. Moore- Collected data through ethnographic research and served as Spanish translator- Accommodated local personal network trends and gender relations through the architectural and social design of a clinic, nursery, and community kitchen on school grounds