Amber Zhang

Member of Technical Staff @ Anthropic

About

I have deep technical expertise in artificial intelligence research, product, and hardware since 2013. AI is not magic. Nor does it automatically make things better. It takes an understanding of the fundamental capabilities and limitations of machine learning models, the human context of the problems to be solved, and the computer systems needed to deliver it at scale. I have what it takes to bring new research to high-quality AI at effective cost, end-to-end. I relentlessly pursue how this technology can do valuable social good instead of polluting our attention with botshit or making a quick buck. Only then can AI perform miracles.

Country

United States

City

Los Angeles Metropolitan Area

Industry

Internet

Skill

Kernel Programming, Concurrent Programming, Operating Systems, Deep Learning, Computational Neuroscience, Convolutional Neural Networks (CNN), CUDA, GPU, C, Programming, Java, Machine Learning, Android Development, Artificial Intelligence (AI), Software Development, HTML, JavaScript, OS X, Linux, Python

Experience

Member of Technical Staff

Anthropic

2026-4 - Present · 6 mos

Senior Software Engineer

Google

2023-4 - 2026-4 · 3 yrs 1 mo

Los Angeles, CA

MLOps for Search and TPU performance optimization for Gemini models.

Software Engineer

Google

2019-2 - 2023-4 · 4 yrs 3 mos

Los Angeles, California, United States

Tech lead for a team of 6 developing NLP infrastructure serving the critical paths of Google Search and Assistant, powering distributed evaluation, releases, experimentation, and debug visualization for 200+ teams. Design APIs offering control over precision–latency tradeoffs, identify and implement concurrent architecture optimizations growing revenue by $millions, and develop a retrieval platform integrating Search with LLMs.

Google.org Fellow

Google

2019-10 - 2020-5 · 8 mos

West Hollywood, California, United States

Consultant for The Trevor Project in artificial intelligence, deep neural networks, and natural language processing. Conceptualized, designed and rapidly prototyped a crisis conversation simulation chatbot, winner of Fast Company's 2021 Innovation by Design Award for Social Good and TIME's Best Inventions of 2021. Trained thousands of counselors and reduced high-risk wait times by 79%. Designed and developed a classifier for assessing LGBTQ youth suicide risk in accordance with a clinical model, with emphasis on AI fairness.

Software Engineer

Google

2015-8 - 2018-2 · 2 yrs 7 mos

San Francisco Bay Area

Developed and maintained cross-platform features for Google Assistant handling millions+ queries per day. Enhanced natural language processing systems including semantic parsing, text generation, dialog management, and i18n. Designed contextual voice UI in collaboration with designers resulting in high user satisfaction. Managed two interns by planning projects and mentoring.

Software Engineering Intern

2014-5 - 2014-8 · 4 mos

San Francisco Bay Area

Implemented distributed recommendation algorithms improving email click-through rate by 2%. Trained and evaluated machine learning models to predict user actions. Created and deployed team-building web application and won best overall project in internal hackathon.

Software Engineering Intern

Stuph, Inc.

2013-7 - 2013-8 · 2 mos

Primary developer of minimum viable product critical to raising $1.5MM in seed funding. Designed and developed website interface with emphasis on usability. Built web scrapers and server-side web APIs.