AI agents and ML models built to do real work.
MS Data Science at UC San Diego · Agentic AI & Machine Learning
At ASML, I build multi-agent AI systems and PyTorch models: a gen AI agent framework that automated 80% of my team's documentation work, and models that predict part performance and failure modes at the design stage. Before that, I worked on genomic data analytics at a biotech startup and portfolio risk analytics in wealth management.
Open to full-time, onsite internships starting January 2027.
Experience
Three roles across semiconductors, biotech, and finance, all built on the same thing: turning data into models people rely on.
-
Jun 2025 – Present · San Diego, CA
AI Engineer Intern · ASML
Developed PyTorch models that predict part performance and failure modes at the design stage, using simulation feedback and hyperparameter tuning. Built a gen AI agent framework with modular workflows that automated most of my team's documentation tasks.
Created a multi-agent blueprint for building end-to-end AI agents, with a cost framework for tokens, tool calls, and execution. One early use case was a risk-assessment agent that reviewed 37 risk events and recommended a vendor for the supply-chain team. I now host open-source LLMs (Qwen, DeepSeek) on vLLM so internal teams can use them securely.
+18% model accuracy80% of docs automated37 risk events reviewed -
Jun 2024 – Sep 2024 · San Diego, CA
Bioengineering Intern, Data Analytics · Aquillius
Ran statistical analysis on genomic datasets in Python and SQL, with feature engineering such as aggregating mutation frequency by gene pathway. Built validation frameworks with stress tests and redundancy checks to model biological system reliability.
+15% model accuracy -
Jun 2023 – Sep 2023 · San Diego, CA
Data Analyst Intern, Financial Analytics · Advanced Trustee Strategy Wealth Management
Analyzed portfolios in Python and SQL, using regression models to find risk exposures and investment opportunities that informed allocation decisions. Built interactive Power BI dashboards that cut manual reporting time.
Regression risk modelsLive Power BI dashboards
Projects & research
Agent systems, LLM infrastructure, neuromorphic research, and robotics.
Agentic AI Blueprint & Evaluation Framework
A modular blueprint that assembles end-to-end AI agents from reusable components, so teams can stand up task-specific agents quickly. Paired with an evaluation framework that scores agents on accuracy, reliability, and task completion.
Open-Source LLM Hosting & Deployment
Hosting and deploying open-source LLMs for an ASML tenant so other teams can access them securely. Qwen and DeepSeek models are deployed and running through an SDS → DIAL → vLLM flow.
Audio Keyword Classification
Built in the Neuromorphic Computing Group under Prof. Jason Eshraghian: a model that classifies spoken keywords by pairing signal processing (MFCC features) with neural networks, reaching 87% accuracy on Google Speech Commands.
Multi-Axis Robotic Arm
A robotic arm with servo motors and sensor feedback for steady, accurate movement, integrating five motor controllers for multi-axis control.
Awards
- Dean's Honor RollUC Santa Cruz · 4.0 GPA · 2023 – 2024
- Eagle ScoutBoy Scouts of America · First inaugural class of female Eagle Scouts · 2021
- 2nd Place OverallUCSD Brain Bee · 2021
Domains
Different industries, the same core work: turn messy data into a model or agent someone can trust.
Agentic AI & LLMs
Multi-agent systems, agent cost and evaluation frameworks, and self-hosted LLM deployment.
Semiconductor design ML
PyTorch models that predict part performance and failure modes from simulation feedback.
Genomic data analytics
Statistical analysis, feature engineering, and reliability validation on genomic datasets.
Financial analytics
Portfolio risk regression and stakeholder dashboards for investment decisions.
Toolkit
Languages
Machine Learning & AI
Data & Analytics
Foundations
Coursework
Education
Let's talk
Based in San Diego. I'm looking for full-time, onsite AI and data science internships starting January 2027, and I'm always happy to talk about agents and ML.