Working on Support Intelligence team to support customer engagement through AI/ML.

• Worked on Alexa shopping team which works to enhance the hands-free (voice-enabled) shopping experience via AI/ML.

• Developed ML dev software for testing and comparing different sentence embedding models within an Alexa speech recognition model architecture.

• Built software to accelerate evaluation process by producing metrics (confusion matrix, accuracy, precision, F1, etc.) in a concise manor, measuring speech recognition architecture performance with different sentence embedders.

• Working on Support Intelligence team.

• Working on project to support customer engagement through AI/ML.

• Worked with Hampshire Hope to collect and analyze data pertaining to the opioid crisis within Western Massachusetts.

• Collected data on transportation options within MA to create heat-maps of MA to visualize price and travel time of possible victims to the closest opioid treatment facility.

• Utilized python to scrape opiod and transportation data from internet, and used libraries such as geoplot and GeoPandas to create heat-maps.

• Responsible for project dealing with reducing the memory usage of NLP machine learning models within company’s servers.

• Built a service based off a web API integrated with an SDK to serve pre-trained NLP model prediction results to other SurveyMonkey services.

• Estimated memory usage reduction of over 70% for NLP models running on server nodes.