Data Scientist · Los Angeles

AI products.
Scientific models.
Evidence at every step.

I’m Rui “Ricky” Sun, a UC Irvine applied mathematics graduate. I build and evaluate AI products and machine-learning systems—from interview preparation and nutrition tracking to conversational data analysis and optical design.

Applied AI · Product Engineering · Scientific ML · Model Evaluation

Open to early-career Data Scientist roles in Los Angeles, Orange County, and hybrid or remote teams.

EducationB.S. Applied & Computational MathematicsUC Irvine · 2026
PracticeAI products + scientific MLModels tied to decisions
ToolkitPython · SQL · PyTorchStatistics + experimentation
LocationLos Angeles, CaliforniaOpen to early-career DS roles

Selected work

From interface to evaluation.

01
AI ProductProduct development

ProofRound

Connecting résumé claims to project evidence, then turning the gaps into focused interview preparation.

Next.jsTypeScriptStructured outputsRealtime voice
Explore ProofRound →
PRODUCT WORKFLOW
01Résumé claim↓
02Project evidence↓
03Interview practice

Problem → implementation → evaluation

02
Health TechnologyNutrition-tracking prototype

Nepholio

A daily nutrition workspace for adults with ADPKD, designed around food records, individual targets, and care-team context.

ReactNutrition dataVoice interfaceCare workflows
Explore Nepholio →
PRODUCT WORKFLOW
01Food record↓
02Daily targets↓
03Care context

Problem → implementation → evaluation

03
Data & AI SystemsCollaborative lab project

Realy

A conversational multi-table analysis workspace that keeps source tables, transformations, and results inspectable.

Next.jsPythonMulti-table reasoningMCP
Explore Realy →
PRODUCT WORKFLOW
01Source tables↓
02Analysis steps↓
03Traceable results

Problem → implementation → evaluation

04
Scientific MLAPS Far West 2026 poster

Generative AI for Nano-Optics

Conditional generation of freeform optical patterns, evaluated with electromagnetic simulation rather than visual similarity alone.

PyTorchConditional VAEDCTRCWA
Explore the research →
SIMULATION WORKFLOW
01Optical target↓
02cVAE candidates↓
03RCWA evaluation

10,000 simulated designs · APS poster

Featured research

Machine learning inside a physics workflow.

My UC Irvine research connects conditional generative modeling with electromagnetic simulation for metasurface inverse design.

Explore the research ↗
END-TO-END OWNERSHIPSCIENTIFIC ML
01Simulation dataGenerate · validate · store
02Conditional modelTrain · compare · diagnose
03Physics evaluationRCWA · refinement · failure analysis
WORKING INSIGHT

The generator is one part of the system. Evaluation design and optimization strategy determine whether an output becomes a useful design.

Experience

From theory to applied decisions.

Full background →
Current work

Scientific Machine Learning Research

UC Irvine · Physics & Astronomy

Building data, model-training, and physics-evaluation workflows for metasurface inverse design.

Portfolio

AI Product Development

ProofRound · Nepholio · Realy

Building user-facing AI workflows with traceable outputs, persistent context, and careful handling of incomplete data.

Jun–Oct 2024

Data Science Intern

Beijing Jinye Paper Co., Ltd.

Worked on forecasting and business analytics in an operating company environment.

How I work

Rigor before polish.

01

Define

Frame the decision, metric, and constraints before choosing a model.

02

Build

Create reproducible data and modeling workflows with clear assumptions.

03

Validate

Use controlled comparisons, diagnostics, and domain-specific evaluation.

04

Communicate

Turn results and limitations into an action someone can defend.

Let’s talk

Looking for a data scientist who can build, test, and explain?

I’m open to early-career Data Scientist opportunities, research collaborations, and thoughtful technical conversations.