Interested in how things work, and how they could work better.

President @ CS + Social Good

Math, CS & Public Policy @ Stanford

Jacob Rubenstein

Turning data into models, software, and decisions. Understanding how real-world systems work and how technology shapes them.

Experience

Currently

CS+Social Good
President of CS + Social Good at Stanford empowering students to apply tech to social impact and build community around public interest tech work

Previously

Citadel Securities
Quantitative trader intern at Citadel Securities built data systems and models that power trading and market making
Stanford
AI and NLP research at Stanford developed coverage-aware retrieval for RAG, including dual residual tracking to cut redundancy without collapsing ranking quality
Recidiviz
Data Engineer Intern at Recidiviz built scalable data pipelines and explored AI to improve criminal justice data and reform outcomes
Stanford Political Science
Researched with Stanford Political Science and Prof. Rob Reich on technology, policy, and governance at the intersection of institutions and society
TreeHacks
Executive leadership for TreeHacks 12,000+ applicants, 1,500+ hackers; led data-driven marketing and logistics for the world's largest collegiate hackathon
US AI Safety Institute
International team intern at the U.S. AI Safety Institute (NIST) developed evaluation frameworks and policy recommendations for frontier AI
Consumer Reports
AI Fellow at Consumer Reports refined RAG pipelines and observability for AskCR, plus state-level AI policy tracking and analysis
Stanford AI Alignment
Analyst at Stanford AI Alignment studied model behavior on alignment benchmarks, evaluation protocols, and approaches to risk mitigation
Deliberative Democracy Lab
Data and ML at the Deliberative Democracy Lab built workflows and analysis for democratic processes, public sentiment, and deliberative outcomes
Cardinal Policy Group
Policy and data research with Cardinal Policy Group for East Palo Alto worked on reparations, economic development, and evidence-based recommendations

Projects & Papers

Coverage-aware retrieval vs standard top-k
CS Research

Diversity-Aware Retrieval for RAG

Dual residual tracking treats retrieval as coverage, not ranking: nDCG stays competitive (0.561 vs 0.612) while redundancy drops from 0.770 to 0.406.

CS Research

ML Clinical Trial Outcome Prediction

Student-teacher framework for spurious correlations; dynamic UnLearning and LLM augmentation.

Math Research

Hausdorff & Gromov-Hausdorff Metrics

Completeness of compact subsets; compactness and isometry under Gromov-Hausdorff distance.

CS Project

PathSense — Hack The North

AI-powered indoor navigation for visually impaired users; voice + CV. First place, Hack the North 2024.

CS Research

Gesture-based Tonal Rendering

CNN + temporal heuristics for real-time guitar strum detection from gestures.

TreeHacks

TreeHacks 2024 Opening Video

Opening ceremony video for 1,500+ participants.

Math Research

Hermite Polynomials

Properties from differential equations; applications in quantum mechanics.

AI Policy

State AI Legislation Tracker

State-level AI bills tracker for policymakers; referenced in state hearings.

Math Research

Secret Sharing Schemes

Shamir vs Blakley; threshold schemes for secure distribution.

CS Research

Trust & Safety in Generative AI

Observability features for AskCR; transparency in generative AI.

TreeHacks

TreeHacks 2025 Opening Video

Directed and edited opening ceremony for 1,500+ participants.

CS Project

Environmental Health Dashboard

Data viz + ML for environmental sensors; health factors in the Philippines.

Math Research

Series Rearrangement Theory

When reordering terms preserves convergence or yields any limit.

CS Research

Unveiling News Biases

ML on news transcripts; political reporting patterns via TF-IDF and clustering.

CS Research

Partisan Sentiment Online

BERT-based analysis of Reddit; topic correlations across partisan discussions.