Longitudinal Risks

• Wrote a position piece advocating for longitudinal studies of human-AI interactions to better understand and mitigate risks that develop over time. Submitted position piece to EMNLP 2026.

Privacy Auditing and Data Provenance

• Improved privacy auditing methods for LLMs to better detect data leakage and inform practical understanding of model risk.
• Proposed a simple and effective canary design that leverages high-temperature sampling to yield highly sensitive synthetic probes for detecting privacy leakage.
• Presented and evaluated an improved approach for auditing synthetically generated data.
• Investigated memorization properties of language models using privacy-auditing techniques.
• Carried out implementation and experimentation.
• Submitted paper to EMNLP 2026.

Models that Forget

• Led a research effort leveraging modularity and unlearning to control data influence and address privacy and policy challenges. 1 of 4 proposals of bold new research directions selected from 56 submissions across the organization.
• Built team of collaborators, drew connections across orgs, charted out research roadmap, secured buy-in from leadership.
• Proposed data compartmentalization as a strategy for AI Governance, presenting a taxonomy mapping governance needs to technical requirements and relevant strategies. Presented paper at GenLaw@ICML 2024.
• Documented internal use cases for unlearning and scoped out partnerships through discussions with product teams.
• Hosted 3 student researchers, investigating approximate unlearning for robustness (preprint) and certifiable unlearning for privacy (preprint).
• Contributed to an internal effort on modular continual learning.

User-level Differential Privacy

• Investigated how to best scale user-level DP training to LLMs under fixed compute budgets.
• Helped define and execute on experiments for the paper.
  

Function-Space Aggregation of Models in Federated Learning

• Led a student researcher project to explore new algorithms for merging models in federated learning, drawing upon ideas in meta-learning.
• Scoped and drafted a proposal; received funding.
• Implemented federated model training and evaluation infrastructure to enable running large-scale experiments on various algorithms.
• Paper accepted to TMLR 2024.

Building a Scalable Dataset Pipeline for Group-Structured Learning

• Developed infrastructure to support scaling up federated learning simulation to datasets and models orders of magnitudes larger than previously used.
• Released an open source library, Dataset Grouper, for partitioning existing flat datasets into group-structured datasets.
• Paper accepted to NeurIPS Benchmarks & Datasets 2023.

Federated Learning AI Explorable

• Published an interactive blog post on how federated learning. Through interactive visualizations the piece allows readers to wrestle with data privacy concerns, learning how FL aims to mitigate them while introducing trade-offs.
  

• Designed a custom compression method for client updates in federated learning to reduce the communication cost from 32 bits per model parameter to 0.1 bits without degrading accuracy on several benchmarks.
• Implemented compression-based aggregation methods for federated learning in TensorFlow Federated. This involved: custom TF ops hosted in TensorFlow Compression, TensorFlow Federated logic, and robust integration tests for system compatibility.
• Presented this work to internal and external audiences. Selected to give an oral presentation at the Google Research Conference. Paper accepted at ICLR 2022 TinyPapers.
• Integrated custom compression method into Google’s federated learning production system.
  

• One of ten graduate students selected to participate in a weekly seminar to learn about the federal policymaking process and develop critical leadership skills to advance science as a public good
• Published an op-ed on a public policy issue; created one-pagers advocating for funding basic scientific research to use in our upcoming congressional visits in Washington, D.C.
  

• Built a deep graph convolutional network (GCN) using Pytorch to predict drug metabolism
• Proposed the use molecular representations learned through GCNs to identify metabolically labile atoms. Compared to traditional feature extraction methods.
• Presented poster at the Rice Data Science Conference, Oct 2019, and the 29th Annual Keck Research Conference, Oct 2019
• Completed written thesis and oral defense in April 2020
  

• Built a data pipeline to query server logs and gather time-series metrics on our services
• Wrote a Spark job in Scala to process and aggregate raw data, storing the results in blob storage
• Developed and implemented an anomaly detection system in Python using Pandas, SciPy and Matplotlib to automatically detect regressions in quality of service among subsets of our network and generate reports to alert iCloud engineers. Deployed system surfaces one to two critical issues each day that otherwise went unnoticed.
• Presented work to ~30 engineers at iCloud and individually to the Vice President of iCloud
  

• Improved an incremental docking protocol (DINC) which computationally predicts how peptides bind to protein receptors. Experimented to identify unexpected behavior; strengthened the robustness of DINC by handling these edge cases.
• Evaluated the latest version of DINC by designing re-docking experiments and writing scripts to automate these tests on the XSEDE Comet Supercomputer. Results published in Devaurs et al, 2019.
• Presented poster at the Rice Undergraduate Research Symposium, April 2018
  

Improving Square’s Appointment Scheduling Calendar

• Optimized the calendar in Square Appointments iOS app by identifying performance bottlenecks and improving the search algorithm. Made a 16-fold improvement in CPU time spent rendering events and UI features that restored calendar to 60 fps scrolling.
• Added a feature to notify users when their time zone differs from that of the business they are viewing

Using Word2Vec to Power a Recommendation Engine

• Developed a customized market insights tool for merchants to compare their prices to those of nearby sellers
• Grouped similar transactions using the “word2vec” ML model
• Built a Python Flask app with D3 Visualization to display interactive reports
  

• Developed an iOS mobile app in Swift that helps users remember the people they’ve met by using location tracking to auto-log events