Thad Hughes

WORK HISTORY

EDUCATION

SKILLS

ABOUT THAD HUGHES

I\'m an engaged, enthusiastic leader and individual contributor (IC) using machine learning, data science, and software engineering to solve problems in a wide range of domains, spanning Natural Language Processing (NLP), computer vision, speech recognition, and the life sciences, including pharma, radiology, and biology. Twelve years at Google followed by three in a biotech startup have developed my rigor, agility, and open-minded approach.At AbCellera, I directed 3 teams (ML, scientific UI, and backend data storage; 16 FTEs total) driving software development for our primary product: high-throughput robotic assays to discover therapeutic antibodies. In addition to directing teams and mentoring team leads, I also directly contributed to R&D, software engineering, data engineering, MLOps, and DevOps, developing ML tools to help scientists understand relationships between antibody sequence space and functional data, resulting in the discovery of antibodies with novel properties and a patent application.As a Tech Lead Manager on Google\'s speech team, I led the “Hey Google” hotword team during Google Home’s launch, trained the first combined “Ok/Hey Google” models and reduced trigger failures by ~40% for accented speakers. As a speech team IC, I single-handedly built both Voice Search in German and Google\'s Android speech data collection app, which we used to crowdsource of hours of speech for acoustic model training. I also dramatically improved noise-robust speech recognition by inventing improved signal processing algorithms, augmenting training with synthetic far-field audio data, and orchestrating the real-world recording of multi-source, multi-microphone datasets.On Google Health\'s radiology team, I curated medical datasets, trained computer vision models to interpret X-rays, and served as a liaison to both internal teams and external clinicians. My work on the speech and radiology teams at Google resulted in several publications, conference presentations, and patents.I love learning, teaching, and communicating, and taught computer science at Munich International School.I\'m fascinated by the connections between ideas, experiences, attention, memories and consciousness, and by our representations of them, including language, music, movement, images, video, maps, plots, math, code, knowledge graphs and other structured data.I care deeply about places and communities at many scales, both urban and natural, and want to help humans thrive while mitigating humanity’s impact on the natural world.