Bridget Salna Email & Phone Number

Cambridge, Massachusetts, US

• Research and development of AI-based solutions to workforce management problems, including:
• Time series forecasting of volume (sales, customers, etc.)
• Predicting risk of employee turnover
• Identifying high-risk shifts for absenteeism/tardiness to drive employee behavior improvement
• Identifying compliance violations in noisy high-dimensional data
• Developing tools to explain ML-based predictions to customers

Boston, MA, US

• Researched, tested, and prototyped models and intelligence for use in production -- specialized in machine learning classification techniques
• Worked cross-functionally with financial experts to build Bayesian networks to model financial strategy recommendations given uncertainty and conditional dependencies in a user's financial profile
• Investigated and prototyped the use of decision and utility theory within a Bayesian network to incorporate user preferences and personalize recommendations

San Francisco, CA, US

• Built AdviceDice (www.advicedice.us), an interactive online application (Flask/AWS) that generates personalized recommendations of new board games based on inputted user preferences
• Collected a data set of ~110,000 ratings from 526 users and features for ~5000 games using web scraping (BeautifulSoup) and XML parsing (ElementTree) of API request data (source: www.boardgamegeek.com)
• Developed model in Python based on user and item-based collaborative filtering of sparse user rating data to predict numerical ratings for new games and make ranked recommendations
• Utilized k-modes clustering to identify similar games based on their mechanics and themes; used these groups to recommend games to users with no previous board game preferences

Boston, MA, US

• Conducted independent research to develop predictive models of an essential biological process (proton transfer) using first principle mathematical derivations, numerical simulations, and nonlinear regression analysis.
• Utilized laser spectroscopy techniques for experimental data collection
• Developed programs in MATLAB to analyze large scale biological data obtained with computational chemistry and apply statistical methods to identify optimal physical characteristics for efficient proton transfer
• Managed and mentored several junior group members to enable them to assist in research project
• Authored four peer reviewed publications

Chestnut Hill, MA, US

• Investigated properties of high temperature superconducting materials using scanning tunneling microscopy experiments
• Constructed program (Fortran) to perform customized numerical Fourier transform analysis of topographic images of material surfaces

Doctor Of Philosophy (Ph.D.), Physics

Master Of Science - Ms, Physics

Bachelor Of Science (B.S.), Physics