Xinqiao Zhang
Senior Machine Learning Scientist at CoreLogic
Specializing in GenAI, Deep Learning, and Trustworthy ML
LinkedIn | Resume | GitHub | Schedule a Meeting
Xinqiao Zhang
Senior Machine Learning Scientist
Specializing in GenAI, Deep Learning, and Trustworthy ML
About Me
Iām a Senior Machine Learning Scientist at CoreLogic, specializing in GenAI and Deep Learning. With a PhD in Computer Engineering from UC San Diego, my research focuses on trustworthy machine learning, AI security, and hardware acceleration for deep learning.
What Others Say
"Xinqiao's expertise in AI security is exceptional. His innovative approach to solving complex challenges and creating robust ML systems has been instrumental to our team's success."
"Working with Xinqiao on hardware acceleration for deep learning models was a game-changer. His methodologies resulted in significant efficiency improvements without compromising accuracy."
Latest News
- š Joined CoreLogic as a Senior Machine Learning Scientist, working on GenAI projects (Oct 2024)
- š Received PhD in Computer Engineering from UC San Diego (Aug 2024)
- š Founded Check-It Analytics, an AI-driven financial information platform (Aug 2023)
What I Bring to the Table
š Problem Solving
I tackle complex ML challenges with innovative approaches, as demonstrated by my work on RAG security that prevented 90% of state-of-the-art poisoning attacks.
š” Innovation Leadership
Led multiple research initiatives resulting in 5 patents and numerous publications in top-tier conferences like ICCV and NeurIPS.
š Business Impact
At Check-It Analytics, I implemented RAG for LLMs to streamline financial processes, achieving up to 80% time savings compared to traditional platforms.
š ļø Technical Depth
Deep expertise in GenAI, ML security, hardware acceleration, and quantization techniques that optimize both performance and security.
Recent Projects
Experience
Industry Experience
Senior Machine Learning Scientist
CoreLogic | Oct 2024 - Present
- Working on cutting-edge GenAI projects
Founder and Chief Technology Officer
Check-It Analytics | Aug 2023 - Present
- Founded and led the development of an AI-driven financial information platform
- Implemented RAG for LLMs to streamline financial processes
- Achieved up to 80% time savings compared to traditional financial platforms
Research Intern
Arm | June 2023 - Sep 2023
- Developed a data distillation algorithm
- Reduced data size to at least 1/10,000 and improved ML model performance by at least 50%
Research Intern
Arm | May 2022 - Aug 2022
- Led a SoC trace data distillation project to improve data processing efficiency
- Developed a novel vulnerability detection algorithm using Graph Neural Networks (GNNs)
- Reduced data processing times by 50%
Academic Experience
Graduate Student Researcher
UC San Diego | Dec 2019 - Aug 2024
- Developed innovative techniques for identifying compromised AI models and enhancing security
- Led a team that achieved 2nd place among 16 competitors in a notable AI security challenge
Education
-
PhD, Computer Engineering
UC San Diego (Advised by Prof. Farinaz Koushanfar) | Aug 2024 -
MS, Electrical Engineering
San Diego State University | Dec 2019 -
BS, Electrical Engineering
Northeastern University (China) | May 2017
Featured Projects
View All Projects ā
Secure Retrieval-Augmented Generation on LLMs
Implemented attack defense using a novel objective function combining adversarial loss, BERTScore, and harmful loss. Successfully prevented over 90% of state-of-the-art poisoning attacks and jailbreaking attacks on RAG-based LLMs.
TrojAI: Detecting Trojans in AI Models
Developed methods to detect adversarial training models. Achieved 2nd place out of 16 teams in the NIST TrojAI competition.
Quantization of Deep Neural Networks
Enabled oblivious inference in Binary Neural Networks. Achieved 2x faster inference and up to 11x faster inference for binary networks.
Collaboration & Impact
"Xinqiao's expertise in AI security and machine learning has been invaluable to our research. His innovative approaches consistently lead to breakthrough results."ā Prof. Farinaz Koushanfar, UC San Diego
Selected Publications
Trustworthy Machine Learning
-
zPROBE: Zero Peek Robustness Checks for Federated Learning
Z. Ghodsi*, M, Javaheripi*, N. Sheyban*, X. Zhang*, K, Huang, & F. Koushanfar
ICCV 2023 | Paper | Patent -
Scalable Binary Neural Network applications in Oblivious Inference
X. Zhang, M. Samragh, S. Hussain, K. Huang, & F. Koushanfar
ACM TECS | Paper