cv

General Information

Full Name Benjamin Ma
Location Kendall Square, Cambridge, MA
Languages English (native), Chinese (conversational)

Education

  • 2022
    Master of Management Science, Global Affairs
    Schwarzman College, Tsinghua University, Beijing, China
    • Guffey Family Fellow in Schwarzman Scholars, a one-year master’s degree and leadership program founded by Steve Schwarzman, CEO and co-founder of Blackstone Group
    • Conducted automated collection, feature extraction, and correlational analysis of top Chinese and Western popular music, and interviewed music industry insiders, to investigate cross-cultural pop music exchange
  • 2021
    M.S., Computer Science | B.S., Computer Science | Minor, Music Production
    University of Southern California, Los Angeles, California
    • Graduated Summa Cum Laude with Discovery Scholar distinction for Outstanding Research
    • USC Viterbi Engineering Fellowship recipient (top 5% of USC engineering students)
    • Studied music production and music industry in Thornton School of Music and Free Artists Matchmakers

Experience

  • 2022
    Software Development Engineer
    Amazon Alexa, Cambridge, MA
    • Managed engineering group responsible for Blitz Hand-Authoring, an Alexa-internal tool for language engineers to create natural language rules guiding Alexa voice interactions
    • Designed system architecture and component interaction model for Blitz Hand-Authoring, including API contracts, NoSQL data storage, AWS Lambda execution environment, and React/Redux front-end
    • Iteratively created Blitz Hand-Authoring front-end UI via hand-crafted Balsamiq wireframes, integrating continuous feedback from Alexa language engineers and localization specialists in the U.S., Europe, and India
    • Performed investigation & audit of high-severity service outage using AWS Cloudwatch metrics, logs, and post-mortem interviews.
    • Conducted full-stack design, implementation, and testing of a highly scalable utterance search engine, achieving a 13x search speedup over baseline by incorporating an Elasticsearch index on AWS DynamoDB
  • 2022
    AI Research Scientist Intern
    Deep Music, Beijing, China
    • Designed and built conditional generation transformer model for simultaneous musical generation of melody and Chinese-language lyrics
    • Adapted mT5 multi-lingual text transformer architecture with custom tokens, callbacks, and loss function to support generation of both text and musical notes; conducted training experiments in Pytorch comparing fine-tuning a pre-trained mT5 model against training from scratch
    • Experimented with different transformations and representation formats for in-house music dataset, and wrote Python scripts to efficiently convert between dense and sparse MIDI representations
  • 2017-2021
    Researcher
    USC Signal Analysis & Interpretation Lab (SAIL), Los Angeles, CA
    • Led SAIL’s music research group investigating characteristic music of major film genres, leading to publication in PLoS ONE, peer-reviewed journal of the Public Library of Science
    • Built multivariate classical and deep machine learning models to forecast human responses to music
    • Wrote Python scripts to automatically clean and collect features for millions of songs from online datasets
    • Achieved top-of-leaderboard and received Distinctive Mention at 2020 MediaEval Emotions & Themes in Music Challenge for the task of building a model to recognize emotion tags for musical clips
    • Developed novel AI interpretability techniques to determine the most relevant musical features for eliciting enjoyment in happy and sad music

Additional Information

  • AWS Certified Solutions Architect - Associate
  • Founder of DJ Lysosome, interactive music production video channel and mobile event DJ service
  • Captain, Schwarzman College and USC Ultimate Frisbee Clubs