Ryan Yee Email & Phone Number

California, United States

San Francisco, California, United States

• Contribute to transforming TV advertising by ensuring data integrity and delivering actionable insights for campaign optimization.• Investigate multiple data sources to identify and resolve systemic data quality issues.• Build and test large datasets to maintain data accuracy and reliability.• Automate and document data processes to ensure best practices in data collection and analysis.• Collaborate with cross-functional teams to analyze TV marketing campaigns and support media buying strategies with impactful reports and visualizations.• Conduct ad-hoc data analysis and deep-dive quantitative analyses to optimize KPIs.• Partner with engineering teams to improve data collection and storage processes.• Support Tatari's mission to make TV advertising as transparent, accessible, and data-driven as digital, empowering brands to make impactful, data-backed decisions.

Cupertino, California, United States

• Developed an AI-powered enterprise knowledge management platform enabling enterprises to interact with data 10x faster, boosting productivity and decision-making.• Built an integrated AI Assistant using Flask, React, NextJS, Tailwind CSS, LangChain, RAG (Retrieval-Augmented Generation), GCP, AWS, and Pinecone VectorDB, providing real-time intelligent responses.• Integrated enterprise learning management systems (LMS) to onboard and assist employees in navigating enterprise documents and manuals, significantly improving administrative efficiency and accuracy.• Engineered a multi-session chat interface with React and Flask, supporting session-specific context and seamless navigation between knowledge packs and chat histories.• Enhanced document accessibility by embedding metadata such as file URLs, names, and relevance scores into vector storage for structured AI-driven document retrieval.• Designed a responsive user interface leveraging ReactMarkdown for rich-text AI responses, including Markdown rendering and embedded citation links for contextual clarity.• Configured Firebase Authentication for secure user management and Firestore for scalable storage of chat sessions, knowledge packs, and associated metadata.• Streamlined vectorization and embedding processes by preprocessing enterprise documents with embeddings and integrating into Pinecone's vector database.• Developed robust API endpoints to handle chat requests, knowledge pack selection, session management, and secure token-based authorization.• Implemented RAG workflows, combining document context with user queries for highly relevant and accurate AI completions.• Optimized fallback mechanisms to gracefully handle insufficient vector matches, ensuring robust performance and reliable AI responses.• Delivered a modular, scalable architecture using a modern tech stack, providing enterprises with a seamless, efficient knowledge management solution.

University Of California San Diego, California, United States

• Collaborated with Marine Scientists to rapidly prototype data preprocessing pipelines, transforming 15 years of sonobuoy data into spectrograms for real-time whale call classification with state-of-the-art machinery to assist in oceanic research.• Developed and deployed an R Shiny App on the official CalCOFI website to provide scientists and non-technical end users with an interactive tool to visualize 20 years of marine mammal data collected onboard CalCOFI.• Integrated multiple data streams, highlighting over 1,000,000 rows of marine mammal visual sightings, eDNA detection, and acoustic detection data through time and space.• Implemented a comprehensive image processing pipeline that addressed challenges in noise reduction through Gaussian filter, bilateral filter, median filter, morphology, and column-wise matrix subtraction.• Conducted Principal Component Analysis (PCA) on spectrograms, resulting in a 45% dimensional reduction while retaining 90% of the total variance. Reduced R-CNN ResNet50 model training time from 10 to 5 minutes per 15 epochs using HPC.• Removed 15 types of vertical artifacts and background noise while preserving essential signals, improving model performance and data usability.• Presented innovative solutions and collaborative initiatives through exhibit booths fostering sustainable ocean observing practices at the Ocean Observing in California joint conference with over 500 attendees.

Cupertino, California, United States

• Conducted data cleaning, data merging, data quality control, and updated dashboard datasets using SQL and Python.• Conducted regression analysis and financial modeling on sales data using R and Excel, optimizing profit margin by 15%.• Used regression analysis and forecasting to define operational KPIs and reduced annual operational costs by 30%.• Compiled 20+ reports in Tableau to provide key stakeholders with actionable information for informed decision-making.• Collaborated with marketing teams to track search engine results page (SERP) and conduct search engine optimization (SEO) using Google Lighthouse, increased organic web search traffic by 200%, inquiry rate by 50%, and sales by 30%.

Santa Clara, California, United States

• Analyzed data storage systems, architectures, and protocols to improve company products: Icebox & Icebox File Manager• Co-authored 10+ blogs on optical storage archives, digital twins, and blockchain technology to promote startup vision.• Conducted market research on data storage products & blockchain technology opportunities related to digital twins.

Bachelor Of Science - Bs, Statistics And Data Science

Student Researcher @ UCSB Data Science Capstone Project

One Year Transfer In Computer Science, Computer Science

High School Diploma, 4.6 Gpa