ML Systems · Data Pipelines · Cloud Infrastructure
Astrophysics student at UC San Diego (Class of 2027) building physics-informed machine learning systems at the intersection of scientific research and real-world infrastructure.
Experience spans end-to-end ML pipelines, from large-scale simulation, data engineering, and Bayesian inference to uncertainty-aware model training and deployment across local and AWS environments.
Currently an Undergraduate Researcher developing deep-learning and probabilistic methods for gravitational-wave population inference, with prior industry experience in cloud, automation, and systems engineering.
Astrophysics student at UC San Diego (Class of 2027) building physics-informed machine learning systems at the intersection of scientific research and real-world infrastructure.
Experience spans end-to-end ML pipelines, from large-scale simulation, data engineering, and Bayesian inference to uncertainty-aware model training and deployment across local and AWS environments.
Currently an Undergraduate Researcher developing deep-learning and probabilistic methods for gravitational-wave population inference, with prior industry experience in cloud, automation, and systems engineering.
Astrophysics student at UC San Diego (Class of 2027) building physics-informed machine learning systems at the intersection of scientific research and real-world infrastructure.
Experience spans end-to-end ML pipelines, from large-scale simulation, data engineering, and Bayesian inference to uncertainty-aware model training and deployment across local and AWS environments.
Currently an Undergraduate Researcher developing deep-learning and probabilistic methods for gravitational-wave population inference, with prior industry experience in cloud, automation, and systems engineering.
Projects
I built a PyTorch-based machine learning pipeline to classify gas morphologies around binary star systems using simulated JWST NIRCam observations. The model consumes 4-channel multi-band images (F200W-F444W) and predicts one of five morphology classes, spiral, shell, bipolar outflow, irregular, or no gas, each corresponding to distinct binary interaction regimes.
Achieved Accuracy
>90%
Macro-F1 across
5 Morphologies
I built a PyTorch-based machine learning pipeline to classify gas morphologies around binary star systems using simulated JWST NIRCam observations. The model consumes 4-channel multi-band images (F200W-F444W) and predicts one of five morphology classes, spiral, shell, bipolar outflow, irregular, or no gas, each corresponding to distinct binary interaction regimes.
Achieved Accuracy
>90%
Macro-F1 across
5 Morphologies
I built a PyTorch-based machine learning pipeline to classify gas morphologies around binary star systems using simulated JWST NIRCam observations. The model consumes 4-channel multi-band images (F200W-F444W) and predicts one of five morphology classes, spiral, shell, bipolar outflow, irregular, or no gas, each corresponding to distinct binary interaction regimes.
Achieved Accuracy
>90%
Macro-F1 across
5 Morphologies
Renewable Energy ROI Analysis Across U.S. States
I led a technical data analysis project evaluating the long-term economic return on investment (ROI) of transitioning U.S. states from fossil-fuel electricity to locally available renewable energy. The pipeline integrates multiple federal datasets and computes three complementary ROI metrics: per-MWh efficiency, total state economic impact, and per-capita equity.
Usability
85%
User Retention
70%
Revamping an E-Commerce Website
Redesigned an existing e-commerce website to improve the user experience and increase sales, including a streamlined checkout process and improved navigation.
Achieved Accuracy
>90%
Macro-F1 across
5 Morphologies
Renewable Energy ROI Analysis Across U.S. States
I led a technical data analysis project evaluating the long-term economic return on investment (ROI) of transitioning U.S. states from fossil-fuel electricity to locally available renewable energy. The pipeline integrates multiple federal datasets and computes three complementary ROI metrics: per-MWh efficiency, total state economic impact, and per-capita equity.
Usability
85%
User Retention
70%
Astronomy Club Newsletter Automation
I designed and implemented a fully automated weekly newsletter pipeline for the UC San Diego Astronomy Club to replace a manual, error-prone communication process. The system aggregates upcoming events from Google Calendar, curates space-science news from multiple sources using LLM-based relevance scoring, and pulls evening-specific weather forecasts optimized for stargazing conditions. It then generates a production-ready HTML email draft in Gmail with subscriber management and a public unsubscribe endpoint.
Conversion Rate
12%
Macro-F1 across
4.8*
Developing a Mobile Health Tracking App
The client, a well-established e-commerce platform, sought to improve their website's navigation to enhance user experience and increase conversion rates.
Achieved Accuracy
150%
Macro-F1 across
5 Morphologies
Astronomy Club Newsletter Automation
I designed and implemented a fully automated weekly newsletter pipeline for the UC San Diego Astronomy Club to replace a manual, error-prone communication process. The system aggregates upcoming events from Google Calendar, curates space-science news from multiple sources using LLM-based relevance scoring, and pulls evening-specific weather forecasts optimized for stargazing conditions. It then generates a production-ready HTML email draft in Gmail with subscriber management and a public unsubscribe endpoint.
Conversion Rate
12%
Macro-F1 across
4.8*
OER Integration at UC San Diego
I am currently working on institutional research and infrastructure planning for Open Educational Resource (OER) adoption at UC San Diego through SPACES. The project focuses on identifying technical, organizational, and policy bottlenecks that limit large-scale adoption, and on designing integration pathways for a centralized OER resource hub serving instructional staff.
Conversion Rate
20%
Macro-F1 across
95%
Optimizing a Corporate Intranet
A leading bank wanted to revamp their mobile app to provide a more user-friendly and secure experience
Achieved Accuracy
>90%
Macro-F1 across
5 Morphologies
OER Integration at UC San Diego
I am currently working on institutional research and infrastructure planning for Open Educational Resource (OER) adoption at UC San Diego through SPACES. The project focuses on identifying technical, organizational, and policy bottlenecks that limit large-scale adoption, and on designing integration pathways for a centralized OER resource hub serving instructional staff.
Conversion Rate
20%
Macro-F1 across
95%
Experience
Experience
Undergraduate Researcher
Undergraduate Researcher
UC San Diego
UC San Diego
Apr 2025 - Present
Apr 2025 - Present
Infrastructure Systems Intern
Infrastructure Systems Intern
General Atomics
General Atomics
Jun 2025 - Aug 2025
Jun 2025 - Aug 2025
Academic Success Program Coordinator
Academic Success Program Coordinator
SPACES at UC San Diego
SPACES at UC San Diego
Jun 2024 - Jun 2025
Jun 2024 - Jun 2025
Resource Coach & Office Clerk
Resource Coach & Office Clerk
OASIS at UC San Diego
OASIS at UC San Diego
Jun 2024 - Sep 2024
Jun 2024 - Sep 2024
Skills
Skills
Python (PyTorch, Pandas. etc)
Deep Learning
ML Algorithms
AWS (EC2, S3, Batch)
Docker
Linux
HPC
SQL
Data Pipelines
Model Evaluation
Automation
Jira
Powershell
Agile Methodology
Leadership
+ More
+ More
+ More