May 2023 – Dec 2025 · 2 yrs 7 mos · San Francisco, California, United States

• Building & leading the internal threat intelligence team at Cloudflare.
• Providing data driven insights to cross-functional partnerships

May 2021 – Sep 2022 · 1 yr 4 mos

• Security Lead:
• Detection & Response (including Security Monitoring Tooling)
• Incident Management
• Threat Assessment
• Hardening

Jun 2018 – Sep 2020 · 2 yrs 3 mos · Mountain View, California

• First successful Sec-ML team at Google.

Jan 2017 – Jun 2018 · 1 yr 5 mos

• Reduced preauthorization failures by 57% by extracting items data from unstructured custom order text fields automatically and computed price estimation for each item saving over $1.5 M per year.
• Built neural network to predict and resolve customer service issues based on SMS messages between couriers and customers.
• Developed metrics for fraud and payments.
• Built NLU model using a graph-like structure to extract item category breakdown of custom order items in raw user-inputted text, giving product the first insight into custom ordered items. This NLU model was also later used to detect fraudulent/ hacked user accounts.
• Designed and developed NLP features for fraud detection such as a name detector and name matching algorithm to parse and model email addresses.
• Built detection service analyzing customer text to mitigate ATO attacks.
• Designed, built, and implemented fraud detection features on distributed systems using a microservice structure.
• Ran and monitored Postgres DB migrations.

Sep 2015 – Dec 2016 · 1 yr 3 mos

• Data Science lead on project in charge of automation of data extraction (built models to extract and parse street address, phone numbers, job data, company descriptions) from tens of thousands of websites in multiple languages to enrich jobs and company data on LinkedIn.
• Built custom graph database to normalize street address detection using phone number area codes, country codes, and street address fragments as signals.
• Working with multiple teams to implement an automatic anomaly detection system using stochastic optimization to improve efficiency of web ingestion.
• Built an algorithm to detect incorrectly labeled profiles based on name of entity, job experience, profile photo analysis (using automatic logo detection to cluster images between logos and photos), and numbers/types of connections.

Feb 2011 – Jun 2015 · 4 yrs 4 mos

• Contributed to several research projects in graph theory, topology, combinatorics, algebra, and data science.
• Led a group of students in an interdisciplinary project over the summer combining graph theory and algebraic techniques to supplement a professor’s paper and supervised their research progress.
• Experimented with computer vision and Robust PCA to classify different human actions in similar settings eg. detecting distracted drivers.