Taro :)
I like data.
Galvanized AI enthusiast. I bridge the gap between raw data and invaluable insights.
Hi, my name is
About Me
I’m a Data Theory student at UCLA with a firm background in statistics and mathematics. Interested in how AI can improve the world, I believe that AI can be a powerful tool to dissolve inequality. I’ll continue to squeeze out every ounce of value from data to help create a brighter tomorrow!
Here are a few technologies I've been working with recently:
- Python
- R
- C++
- SQL
- PyTorch/Tensorflow
- Transformers
- Scikit-learn
- Bash
- HTML/CSS
- Django
Experience
- Improved pipeline performance by 3x by integrating CUDA parallel GPU processing and unifying embeddings between pipeline components
- Fine-tuned and experimented with SOTA fact verification models like DeBERTa and MultiVERs
- Built a multi-source evidence retrieval system with Pinecone optimized with multi-threading, integrating it into a claim verification pipeline
- Developed and dockerized evaluation scripts using FEVER, SciFact, and HealthVER datasets for automated testing workflows with Github Actions
Truth Systems seeks to enable the usage of LLMs in the medical space by eliminating hallucinations with a state-of-the-art AI-powered fact verification pipeline.
- Led three teams through the computer vision MLOps cycle, including data collection with SerpAPI, training CNN and ViT models for classification and object detection, monitoring metrics with TensorBoard/Weight and Biases, and deployment with Flask.
- Cut hallucination rates by 50% in a GPT-powered Q/A application by leveraging LangChain and the ReAct (Reasoning and Action) framework for RAG.
- Utilized multiple tools such as vector databases (Chroma), calculators, Wikipedia, etc. to drastically increase the capabilities of LLMs.
AI Camp is revolutionizing how we see AI, developing an entourage of 1000+ pre-College students equipped with the foundational knowledge and experience necessary to bring upon further AI advancements.
- Leveraged various growth strategies and tools including social media, email marketing, presentations, and peer and faculty member networking to grow the user base
- Strategically assessed growth and performance metrics to improve, change and/or help design new growth strategies
RippleMatch is the recruitment automation platform changing how Gen Z finds work. By replacing job boards with matching and automation, RippleMatch eliminates the most time-intensive parts of the recruitment process for both employers and job seekers. Leading employers such as Amazon, eBay, and Teach For America leverage RippleMatch to build diverse, high-performing teams and Gen Z job seekers across the country trust RippleMatch to launch and grow their careers.
- Provided 100+ under-served patients with vital Covid vaccines
- Compiled 100+ patients’ data into Aviva Pharamcy’s database
- Optimized patients’ benefits based on their insurance options
- Created 10+ professional relationships between Aviva and local businesses
- Scheduled 10+ vaccination appointments for new patients
- Assisted senior citizens with online vaccine application process
During the midst of the Covid-19 Pandemic, Aviva Pharmacy provided Pfizer and Moderna’s Covid-19 vaccines to people throughout the Greater Los Angeles area. As the vaccines were limited to essential workers and people over the age of 65, Aviva’s goal was to help eligible people obtain them as quickly and easily as possible. With a limited staff, their Covid-19 vaccination distribution program was heavily dependent on volunteers.
- Sourced 300+ candidates through LinkedIn Recruiter
- Cleaned and analyzed data with Excel by merging datasets, creating new variables through existing ones, and visualizing candidate statistics.
- Transferred data visualizations onto a website using the SquareSpace Interface
- Created an internal promotional video using Adobe Premiere Pro
Ejento (formerly TS2) is a strategic business development firm for the technology industry. Ejento is a unique firm that advises technology startups and connects them to investors and technology professions allowing them access to the resources they need to thrive. With experienced experts in gaming, AR/VR, adtech/martech, e-commerce, and mobile technologies, the Ejento team works in the Silicon Beach area to help tech companies throughout the startup universe.
Projects
Learning Equality; Curriculum Recommendations
Using Siamese Networks and Natural Language Processing to match educational topics with multiple materials (multi-label).
Predicting Car Accident Severity
Using randomForests to predict the severity of car accidents in the United States.
Student Performance Analysis
Analyzing relationships between educational and socioeconomic factors and how they affect student performance.
Opioid Crisis and Possible Resolutions
Given drug use data of several US states, we created a mathematical model that predicts the regions where opioid use originated and spread.
Polar vs Non-Polar Bonds
A short video describing the differences between polar and non-polar bonds... in Minecraft.
Education
2020 - 2024
Bachelor of Science in Data Theory
University of California, Los Angeles
GPA:
3.85
Relevant Coursework:
- Computational Statistics
- Object Oriented Programming
- Numerical methods
- Regular Expressions
- Webscraping
- TidyVerse Data Visualization/Manipulation
- Randomization/Bootstrap Tests
- Python and Other Technologies for Data Science
- Pandas, Scikit-learn, SQL, Git
- Intro to Statistical Models and Data Mining
- Regression and Classification Analysis (linear, logistic, kNN, LDA/QDA, Ridge/Lasso, PCR, SVM, GAM, Boosting, randomForests, Artificial Neural Networks, XGBoost)
- Cross validation
- Feature Engineering (backward/forward feature selection, lasso/ridge regression, PCA)
- Imputing missing values (MICE, Amelia, kNN Imputation, missForest, Hmisc, rpart)
- Unsupervised Learning (K-Means Clustering, Hierarchichal Clustering)
Get in Touch!
My inbox is always open. Whether you have a question or just want to say hi, I’ll try my best to get back to you!