Megan Barnes Email and Phone Number

Intern at Speech Technology & Research Lab, working on multi-modal (text and speech) modeling for question identification.

• Built features within a custom LSTM neural net architecture (using Torch) for text classification and entity recognition. Productized state-of-the-art machine learning research and innovated novel techniques.• Created benchmarking scheme and datasets for measuring improvement in the product’s machine learning technology• Implemented new back-end features and maintained a high-availability cluster with >4,500 users• Designed and implemented a full overhaul of the product’s internal RESTful API

• Designed research questions and identified viable data sources, methods, and metrics for analysis• Implemented data aggregation pipelines: scraping, cleaning, normalization, and statistical analysis• Interpreted research and drew meaningful conclusions from statistical results for use in academic papers

• Increased usability of a high-traffic, large-codebase website (software.intel.com) in a development team setting• Built new webpage features, implementing both front- and back-end code