Marvin Li Email and Phone Number
Marvin Li work email
- Valid
- Valid
Marvin Li personal email
I am an undergraduate at Harvard studying computer science and math. Outside of class, I am a machine learning researcher at the Harvard SEAS. I am interested in the science of deep learning and the privacy and robustness of language models.With Prof. Sitan Chen at Harvard SEAS, I helped discover and prove a theoretical framework for feature emergence in diffusion models. In my research with Prof. Seth Neel at the Harvard Business School, I investigated the privacy risks of large language models with gradient-based attacks.
Harvard John A. Paulson School Of Engineering And Applied Sciences
View-
Research Assistant For Prof. Cynthia DworkHarvard John A. Paulson School Of Engineering And Applied SciencesSalisbury, Md, Us
-
Research Assistant For Prof. Cynthia DworkHarvard John A. Paulson School Of Engineering And Applied Sciences Apr 2024 - PresentCambridge, Ma, Us- Theoretically and empirically analyzed how geometric bias in prompt embeddings leads to representational bias in text-to-image diffusion models for race and gender (ICML 2024 TiFa Workshop).
-
Research Assistant For Prof. Sitan ChenHarvard John A. Paulson School Of Engineering And Applied Sciences Oct 2023 - PresentCambridge, Ma, Us- Developed a theoretical framework to explain feature emergence in diffusion models for image generation. - Applied framework to audit privacy and fairness violations for Stable Diffusion 2.1. - Published first-author paper at ICML 2024. - Gave invited talk at the Donder's Institute for Brain, Cognition, and Behaviour (https://www.youtube.com/watch?v=NbnBmi5k1UQ).
-
Research Assistant For Prof. Seth NeelHarvard Business School Feb 2023 - PresentBoston, Ma, Us- Developed MOPE, a novel zero-shot privacy attack against LLMs that identifies individual samples from pre-training data by exploiting gradient information.- Developed a loss ratio attack which extracts >90% of LLAMA-7B's finetuning dataset. - Published co-first-author paper at EMNLP 2023. - Contributed to a comprehensive codebase and library to evaluate different privacy attacks against LLMs. -
Algorithm Developer Intern (Quantitative Research Intern)Hudson River Trading May 2024 - Aug 2024New York, Ny, UsInterned on the international equities and options teams. -
Quantitative Research InternCitadel Securities Jun 2023 - Aug 2023Miami, Florida, UsInterned on the Systematic Equities team. -
Software Engineering InternTwo Sigma Mar 2022 - Aug 2022New York, Ny, Us• Created a manual order workflow tool in Java/SQL to liquidate out-of-universe securities.• Wrote the backend API endpoints of an internal web tool for users to search, create, and edit manual orders. • Used jOOQ for SQL queries in Java, JUnit and Mockito for unit‑testing, and Swagger for API management. -
First-Year Trading & Technology Program (Fttp)Jane Street Mar 2022 - Mar 2022New York, Ny, UsSelected for Jane Street’s 3-day program; Placed 2nd in Electronic Trading Competition (ETC). -
Research InternUniversity Of Maryland Center For Environmental Science Aug 2018 - Jan 2021Cambridge, Maryland, Us• Analyzed in-situ Karenia brevis sampling data, streamflow data from the United States Geological Survey, satellite altimetry data, and wind buoy data from the National Data Buoy Center (in R).• Developed machine learning models to predict outbreaks of Harmful Algal Blooms on the West Florida Shelf.• Used machine learning to investigate how specific physical and biological processes, climate change and nutrient management affect the frequency of K. brevis blooms.• Extended the machine learning models to global oceans. -
Research InternUniversity Of Maryland Center For Environmental Science Jul 2016 - Nov 2018Cambridge, Maryland, Us• Analyzed remote sensing reflectance data from the MODIS Aqua satellite as well as in-situ total suspended sediment (TSS) and chlorophyll data (in R).• Developed ensemble machine learning models (relevance vector machines, support vector regressions and artificial neural networks) and generalized linear models to infer suspended sediment and chlorophyll concentration from MODIS Aqua satellite reflectance data (in R).
-
ScholarResearch Science Institute (Rsi) Jun 2020 - Aug 2020Cambridge, Massachusetts, Us• Developed machine learning algorithms to predict catalyst efficiency from properties• Used Bayesian optimizations to tune catalyst parameters• Found several novel catalysts which outperformed conventional ones• Verified them with Density Functional Theory
Marvin Li Skills
Marvin Li Education Details
-
Harvard UniversityMath And Computer Science -
Massachusetts Institute Of TechnologyComputer Science
Frequently Asked Questions about Marvin Li
What company does Marvin Li work for?
Marvin Li works for Harvard John A. Paulson School Of Engineering And Applied Sciences
What is Marvin Li's role at the current company?
Marvin Li's current role is Research Assistant for Prof. Cynthia Dwork.
What is Marvin Li's email address?
Marvin Li's email address is ma****@****gma.com
What schools did Marvin Li attend?
Marvin Li attended Harvard University, Massachusetts Institute Of Technology.
What skills is Marvin Li known for?
Marvin Li has skills like Programming Languages, Regression Analysis, Machine Learning, Statistical Data Analysis, Laboratory Skills.
Free Chrome Extension
Find emails, phones & company data instantly
Aero Online
Your AI prospecting assistant
Select data to include:
0 records × $0.02 per record
Download 750 million emails and 100 million phone numbers
Access emails and phone numbers of over 750 million business users. Instantly download verified profiles using 20+ filters, including location, job title, company, function, and industry.
Start your free trial