# Andy Zhao

> AI Research Engineer | MIT&Berkeley Alum | LLM Startup

Location: Mountain View, California, United States
Profile: https://flows.cv/andyzhao

As a data scientist at Ameriprise Financial Services, LLC, I build predictive models to forecast geographical location growth for real estate investment in Europe. I employ deep neural networks and boosting algorithms to improve model performance and interpretability, and develop data-specific training metrics advised by Professor Rama Ramakrishnan. Previously, I worked as a machine learning engineer at the United Nations Mine Action Service, where I built and evaluated models to identify new buildings, roads, human settlements, and land cover on satellite imagery, and to assess the effects of mine clearance projects in Colombia. My team won the second place in the case competition out of 40 teams.

I am also a graduate student in optimization at MIT, where I research on theoretical computer science and mixed integer programming under the direction of Professor Dimitris Bertsimas. I propose and implement algorithms that work better than the Gurobi Solver on some problems, and develop a machine learning scheme to maximize cancer patient survival time for various treatments using optimal policy trees. I have a strong background in computer science, mathematics, statistics, and data science, having completed a quadruple major at Berkeley, and having participated in research projects at Tsinghua University and Stanford University. I am passionate about exploring large language models and their applications, and I have published papers on quantum random number generation and image processing. I am fluent in Chinese and English. I am looking for a job in data science or data analysis that challenges me to use my skills and knowledge to solve real-world problems.

## Work Experience
### Gen AI Research Scientist @ Samsung Research America (SRA)
Jan 2025 – Present | Mountain View, California, United States
Research and implement state-of-the-art research:
1. Optimize hallucination of internal LLM, and LLM Evaluation
2. Build recommendation system of advertisement bidding using LLM reasoning

### Founding Engineer @ Stealth AI Startup
Jan 2023 – Present | San Francisco Bay Area
Large Generative Model Engineer/Research Scientist

### Research Assistant @ MIT Computer Science and Artificial Intelligence Laboratory (CSAIL)
Jan 2022 – Jan 2023
• Proposed and implemented polynomial-time approximation algorithms for NP-hard problems
• Proposed optimize to learn simplex algorithms, a online heuristic optimizer analogy to Adams

### Data Scientist @ Ameriprise Financial Services, LLC
Jan 2023 – Jan 2023 | Boston, Massachusetts, United States
• Built predictive models to forecast geographical location growth in Europe for real estate investment
• Researched and developed data-specific training metrics to enhance model performance

### Applied Scientist Co-op @ United Nations
Jan 2022 – Jan 2022
• Built models to identify new buildings, roads, human settlements, and land cover on satellite imagery
• Evaluated effects of Columbian mine clearance projects by CNN, Transformers (2nd place in contest)

### Quantitative Research Intern @ Alpha Square Group
Jan 2022 – Jan 2022 | New York City Metropolitan Area
• Created NLP model to: web-scrap alternative time, summarize reviews, and evaluate sentiment
• Constructed evaluation system for primary market investment based on historical investment

### Research Assistant @ Berkeley Artificial Intelligence Research
Jan 2020 – Jan 2022 | Berkeley, California, United States
1. Developed and benchmarked model-update compression methods for federated learning
2. Developed CNN based methods for arbitrary fold of symmetry in nuclear reactor image

### Machine Learning Engineer @ Goodly Labs
Jan 2019 – Jan 2020
• Created new initiatives such as the Public Editor FAQ, Triage Extension, and Website Redesign 
• Performed sentiment analysis and causal inference for misinformation detection and evaluation

### Software Engineer Intern @ Tencent
Jan 2018 – Jan 2018 | Beijing
• Developed a conference room management system utilizing Constraint Satisfaction Problems algorithms to optimize resource allocation


## Education
### Doctor of Philosophy - PhD in Electrical Engineering and Computer Science
University of Southern California

### Master's degree in Operation Research and Applied Math
Massachusetts Institute of Technology

### Bachelor's degree in Quadruple Major: Computer Science, Pure Math, Statistics, Data Science
University of California, Berkeley

### Computer Science
Stanford University

### Research in Institute for Interdisciplinary Information Sciences
Tsinghua University

### High School Diploma in Mathematics
PRINCETON INTERNATIONAL SCHOOL OF MATHEMATICS AND SCIENCE, INC.

### Middle School Diploma in Mathematics
The High School Affiliated to Renmin University of China


## Contact & Social
- LinkedIn: https://linkedin.com/in/zehao-zhao
- Portfolio: https://scholar.google.com/citations?user=g7p2U-wAAAAJ&hl=zh-CN

---
Source: https://flows.cv/andyzhao
JSON Resume: https://flows.cv/andyzhao/resume.json
Last updated: 2026-04-11