Built and evaluated multiple unsupervised anomaly detection models (Prophet + Isolation Forest, Merlion, LSTM-AE) for API telemetry monitoring; selected and deployed Prophet-based pipeline for best performance, interpretability, and ease of deployment.
Applied AWS Bedrock's Titan Embeddings and Claude 3.5 Sonnet for clustering, naming, and sentiment analysis of NPS feedback, extracting key themes to support product insights from unstructured text.
Conducted keyword analysis on AI Advisor open-field queries to assess reference URL coverage, uncover content gaps, and improve user-facing query resolution.
Developed Looker dashboards to track anomaly alerts and AI-driven insights, enhancing visibility across teams and automating Slack-based reporting to reduce manual monitoring efforts.
Pinecone
Data Science Intern
June 2024 – August 2024 · New York, NY
Designed and built the Book of Business and Account 360 dashboards using SQL and Sigma, improving sales operations by 15%. Implemented Row-Level Security (RLS) for tailored views, and documented processes in Notion, reducing onboarding time by 30% and ensuring consistent use across teams.
Developed the "dim_assistants" schema and implemented it in the pipeline using BigQuery and DBT. Created the Pinecone Assistant dashboard using SQL and Sigma, enabling comprehensive tracking of metrics. Facilitated cross-team collaboration, leading to a 25% increase in product insights.
Conducted churn analysis using Python and Random Forests, identifying 5 key metrics and setting up alerts, reducing churn by 10%. Overcame data limitations and improved data collection, projected to boost accuracy by 20%.
Allschool Inc.
Data Analyst Intern
June 2022 – August 2022 · San Mateo, CA
Enhanced impression targeting strategies and boosted customer engagement through A/B testing and segmentation analysis of user traffic and revenue across regional and platform data utilizing Google Analytics.
Evaluated user behavior across multiple advertising channels, leading to a 15% reduction in project budget and an increase in daily active users employing SQL and BI tools.
Designed a real-time web scraper with Python and Selenium, accelerating the class selection process by 50%.
Developed a key metrics dashboard for active users, daily traffic, and revenue, improving business visibility and supporting data-driven decision making using Google Looker Studio.
UC Davis Economics
Research Assistant
July 2020 – September 2020 · Davis, CA
Analyzed behavioral trends in procrastination and present-biased behavior using a generalized linear model (GLM) and Logistic Regression.
Enhanced the reliability of study results by employing Bootstrapping resampling techniques to expand the sample size to approximately 20,000 data points.
Addressed multicollinearity among predictors and improved prediction accuracy of procrastination behavior variables with regularization using Lasso Regression.
Facilitated industry application by uncovering procrastination patterns, offering insights for tech companies to develop user-centric products and services, potentially enhancing user satisfaction, retention, and success.
Launchpad Project Management
Data Analyst Intern
June 2019 – September 2019 · Davis, CA
Utilized R and SQL for comprehensive market analysis and database management across multiple portfolios, including enabling the development of a predictive model for real estate investment strategies.
Applied machine learning algorithms and statistical methods like generalized linear regression and logistic regression to analyze survey data, extracting critical insights for the portfolios.
Employed advanced data visualization tools like Tableau to present data analysis results, facilitating strategic decision-making and enhancing team understanding.
