Columbia University

Master’s of Science in Computer Science (Machine Learning Track) – Sep 2025 – December 2026

Quantum Coursework:

• COMS6998E Topics in Computer Science: Quantum Error Correction
• EECS6890E Topics-Information Processing: Quantum Engineering
• PHYS5084GR Quantum Simulation and Computing Lab
• ELEN4730E Quantum Optimization and Machine Learning
• COMS4281W Introduction to Quantum Computing

Other Relevant Coursework:

• COMS6997E Topics in Computer Science: Machine Assisted Math
• COMS4776E Neural Networks and Deep Learning
• COMS4762W Machine Learning for Functional Genomics

Teaching Assistant

• COMS4281W Introduction to Quantum Computing

Danish Institute for Study Abroad Copenhagen

Jan 2024 – May 2024

Relevant Coursework:

• Artificial Neural Networks and Deep Learning
• Artificial Intelligence

Israel Summer Business Academy

Jun 2022 – May 2022

• MGT 200 Venture Creation
• INTL 320 Business, Innovation and Entrepreneurship

Menlo School

High school Diploma – Aug 2017 – May 2021

• 4 years of Computer Science
• 5 on AP Computer Science A

Cidon Systems Research Lab

Graduate Student Researcher – Columbia – Sep 2025 – Present

Collaborating with Dr. Asaf Cidon and Barracuda Networks, I explore machine-learning approaches to detect spam and phishing attacks. My work involves developing and evaluating new classification pipelines that combine language-model embeddings with real-world email telemetry to improve detection precision while minimizing false positives.

Desdr Open Insurance Toolkit

Graduate Student Researcher – Columbia – Sep 2025 – Dec 2025

Link

Working with Dr. Eugene Wu on building an open-source insurance modeling toolkit for NGOs and governments serving rural farmers in low-data regions. Our team collects field data through SMS-based surveys and converts it into actuarial and climate-risk features. I help design models and dashboards that estimate expected losses and guide organizations in creating affordable micro-insurance programs.

Foster System Database

Undergraduate Student Researcher – WashU – Sep 2024 – June 2025

Github link

Under the supervision of Dr. Ian Fillmore, I fine-tuned a LLaMA 3 model using Unsloth to answer detailed queries about foster-care policy data. I implemented Human-in-the-Loop reinforcement learning on Hugging Face Spaces to iteratively refine model responses. This system enabled cross-state policy comparisons and provided researchers with an interactive way to analyze foster-care outcomes.

Highnote

Cybersecurity Engineer Intern – Summer 2024

At Highnote, a fintech startup focused on card issuing and processing, I took ownership of the company’s Security Information and Event Management (SIEM) system. I unified logs, metrics, and alerts from AWS, GCP, and Datadog into a single Elasticsearch instance, improving visibility and simplifying incident response. I configured every component of the SIEM—agents, integrations, pipelines, data streams, indices, and APIs—to support more than a terabyte of daily log ingestion. I implemented over 100 alerts for suspicious activity such as logins, commands, and downloads, combining hard-coded rules with new machine-learning and cross-system detections to strengthen security monitoring.

Bugcrowd

Associate Application Security Engineer – Summer 2020

At Bugcrowd, a global bug-bounty and vulnerability-management platform, I managed end-to-end bug triage operations, reviewing and validating vulnerability submissions from a worldwide network of researchers. I collaborated with Fortune 500 and major technology clients to confirm findings, prioritize remediation, and implement long-term preventive measures. This work sharpened my understanding of coordinated disclosure workflows and scalable security program management.

University Club of Palo Alto

Lifeguard – Jun 2018 – Jan 2020

At the University Club of Palo Alto, I cultivated leadership and crisis-management skills while ensuring member safety. I managed crowded pool environments, enforced safety protocols, and responded calmly to emergencies. This role taught me composure, quick decision-making, and teamwork under pressure.

WhatTheDuck iQuHacks 2026

Github link

For the iQuHACK 2026 State Street × Classiq challenge, I designed the classical Quasi-Monte Carlo baseline and contributed to the design of the quantum Iterative Amplitude Estimation (IQAE) approach for Value at Risk (VaR) estimation. I implemented and benchmarked both pipelines using CUDA-Q, including reusable state preparation, threshold oracles, and a bisection-based VaR search, enabling large-scale and fair evaluation. I ensured rigor by enforcing identical loss models, discretization schemes, confidence levels, and stopping criteria across classical and quantum methods, empirically validating the expected O(1/ε²) versus O(1/ε) scaling behavior. Our work received an Honorable Mention in the State Street × Classiq category and won the NVIDIA Ecosystem Award for GPU-accelerated quantum–classical benchmarking.

Practical LLM-Based Lossless Compression

Github + paper link

This project explores the practicality of using large language models for lossless text compression by combining autoregressive LLMs with arithmetic coding. We implemented an optimized compression pipeline and evaluated multiple open-source models, finding that LLMs achieve 5–16× average compression and up to 39× on individual files, significantly outperforming classical compressors like gzip and zstd. I focused on systems-level optimizations, including static KV caching, CUDA graph capture, and batched segment processing, and analyzed the trade-offs between compression ratio, throughput, and GPU memory usage.

ASLingo

Github link

As part of my Software Engineering Workshop, my team built ASLingo, a Duolingo-style web app for learning American Sign Language. I led the computer-vision component, developing a real-time hand-sign translation pipeline using Google’s MediaPipe API and a RandomForestClassifier for rapid inference. I later integrated a Long Short-Term Memory (LSTM) model to process temporal sequences in video, enabling accurate continuous-sign recognition.

In progress:

Columbia Autonomous Racing Club

Collaborating in the first ever semester of the autonomous racing club at Columbia. We are fundraising with companies and adapting another member’s past project from another university.

WashU Votes

Through WashU Votes, I promoted civic engagement across campus and the broader St. Louis community. I helped organize voter-registration drives, managed polling booths, and led outreach events to raise awareness about election updates and voting-law changes. I also represented WashU at the Missouri State Capitol in Jefferson City, meeting with legislators to advocate for accessible voting rights.

WashU Gymnastics Club

Gymnastics has been one of my longest-running passions. After years training with Stanford Boys Gymnastics, I joined the WashU team, where I led practices and compete at collegiate meets. The shift from a youth program to a student-run club taught me leadership, resilience, and collaboration in a new competitive setting.

Biking

I love to bike tour! All of my trips I’ve carried tent, food, and all supplies on our bikes with no van support.

Biggest bikes:

• 3,200 miles from Savannah, Georgia to LA, California – Summer 2019
• 1,500 miles from Amsterdam, Netherlands to Barcelona, Spain- Summer 2018
• 1,000 miles in Canadian and Montanan Rockies – Summer 2021
• 2025 Death Ride, 115 miles with 14,000 feet of elevation through Sierra Mountains – Summer 2025
• 400 miles from Jasper(!), Canada to Banff, Canada – Summer 2025

Ironman

• 2025 Oceanside Ironman 70.3: 5:18
• 2025 Lake Placid Ironman: Stress Fracture /: