**Yao Huang (She/Her)**
(217) 418-4396 | San Francisco, CA
yao_huang@berkeley.edu | www.linkedin.com/in/yao-huang-yoshi-feeder

A self-motivated software engineer passionate about the intersection of machine learning and digital accessibility.

### **SKILLS**
**Languages:** Java, Python, SQL, CSS, HTML, JavaScript, C++, MATLAB, R
**Tools & Frameworks:** Linux, Jira, Git, Docker, Spring MVC, PyTorch, TensorFlow, Keras, Pandas, NumPy, Scikit-learn, AWS RDS, AWS EC2, MySQL, Redis Cache, RESTful API, AJAX, Jenkins, Scrum, MonkeyLearn API, Neo4j
**AI Models:** Transformers, YOLOv4, Diff GAN, FastSpeech 2, HiFi-GAN, Tacotron2

### **PROFESSIONAL EXPERIENCE**
**Software Engineer Intern, Metafoodx**
*02/2024 – 04/2024*
- Developed Python Open3D point cloud visualization tools, improving low-volume estimation accuracy by 80%.
- Created an automated image extraction optimization tool, increasing food image embedding population efficiency by 15x.
- Root-caused and fixed IoT logging and image warping issues blocking production release.
- Improved AI classification of similar-looking food items and trained model with new images, achieving 98% accuracy.

**Software Engineer Intern, E Shopper Fulfillment LLC**
*04/2023 – 01/2024*
- Developed an e-commerce site with search, favorites, and product recommendations, increasing user visit time by 30%.
- Built Java Servlets for login, session, search, and recommendation features, using encryption for security.
- Designed a content-based recommendation engine and optimized response time by 90% using Redis Cache on AWS EC2.
- Constructed frontend UI with JavaScript, HTML, CSS, and AJAX for seamless user interaction.

**Software Engineer Intern, Bo Rui Tong Software**
*06/2020 – 08/2020*
- Developed a Neo4j graph database tool to detect fraud, reducing suspicious transactions by 20%.
- Implemented an AI-based classification tool automating back-office operations with 98.62% accuracy.

### **CAPSTONE RESEARCH PROJECTS**
**Neural Network for Hand Gesture Classification**
*08/2021 – 05/2022 (PI: Prof. Brian A. Barsky)*
- Curated a 30,000+ image dataset for YOLOv4 training, improving multi-object detection accuracy to 90.9%.

**Sign Language Recognition – Text-to-Speech (TTS)**
*08/2022 – 12/2022 (PI: Prof. Brian A. Barsky)*
- Trained Tacotron2 and HiFi-GAN models to convert American Sign Language to human-like speech.

### **EDUCATION**
**University of California, Berkeley**
*Master of Engineering in Computer Science | GPA: 3.9/4.0*

**University of Illinois Urbana-Champaign**
*Bachelor of Science in Mathematics, Statistics, and Economics (Triple Major) | GPA: 3.9/4.0*
**Honors:** Phi Beta Kappa, Phi Kappa Phi, Dean’s List (Fall 2014 – Spring 2017)

no comments

✔ list