Available for full-time roles · May 2026

Hi, I'm Carina Sun.

Data Scientist &
AI Engineer &
From Psychology.

I specialize in agentic AI systems, LLM pipelines, and scalable analytics — blending behavioral science with applied machine learning to build human-centered AI.

Carina Sun

Where Data Meets Human Insight

I'm a Data Scientist & AI Engineer completing my Master's in Data Science at NYU, with a background in rigorous quantitative methods and the psychological nuances of human behavior.

My work spans production ETL systems at Amazon, multi-agent AI architectures at the United Nations, and ML research on large language models — always with an emphasis on building systems that are not just powerful, but purposeful.

My psychology background gives me an edge in designing models that account for behavioral patterns, not just data patterns.

New York University
Master of Science in Data Science
Machine Learning · Deep Learning · NLP · Causal Inference · Big Data Systems
Expected May 2026
New York University
BA in Psychology (Honors) · Minor in Math & CS · Business Studies
GPA 3.85 · Presidential Honors Scholar · Dean's List · Dean's Research Fund
Experience at
Amazon
United Nations
NYU

Built to Build

🐍
Languages
Python (Expert)SQL (Advanced)RJavaJavaScriptMATLABGitDocker
🤖
Machine Learning & AI
PyTorchScikit-learnXGBoostA/B TestingPandasNumPy
🧠
Generative AI & LLMs
LangChainLangGraphGPT-4oLLaMA-3.1RAG ArchitecturePrompt Engineering
☁️
Data Engineering
AWS (S3, Redshift)SparkTableauPower BIETL PipelinesHadoop
📊
Statistical Models
Causal InferenceBehavioral ModelingPsychometricsExperimental DesignRegression AnalysisComputational Modeling
🎤
Teaching & Communication
Graduate ML InstructionTechnical MentorshipExecutive PresentationsCurriculum DesignCross-functional Collab

Where I've Made Impact

May – Aug 2025
New York, NY
Amazon
Business Intelligence Engineer
Built production ETL pipelines and predictive models to automate supplier evaluations, drive executive strategy, and scale platform capacity.
↑65%Data Velocity
↓90%+Entry Errors
Platform Scale
500+Eng. Hrs Saved
Sep 2025 – Present
New York, NY
United Nations
AI Engineer, Capstone
Architecting a modular multi-agent AI system with LangGraph to automate data analysis and anomaly detection across international procurement records.
50K+Records
~85%Anomaly Precision
↓60%Debug Time
200+Analyst Hrs Saved
Jun – Aug 2023
New York, NY
Omnicom Media Group
Marketing Data Analyst
Designed ETL pipelines and Tableau dashboards to centralize ad performance data, and ran A/B tests on media buy strategies to optimize campaign ROI.
↑12%ROI Improvement

Classes I Tutor

🤖
NYU · Graduate Level
Introduction to Machine Learning
Teaching Assistant for a graduate-level ML course, supporting 100+ students in understanding Neural Networks, SVMs, and real-world model design, evaluation, and debugging with PyTorch and scikit-learn.
Jan 2026 – Present
🧪
NYU · Undergraduate
Statistics & Data Analysis for Research in Psychology
Supporting psychology students in applying rigorous statistical methods to research design — bridging quantitative analysis with behavioral science for real-world research applications.
Sep 2024 - Sep 2025
👩‍💻
NYU Courant · GSTEM Program
Data Science & Machine Learning for High School Girls
Teaching assistant for GSTEM — NYU Courant's flagship program empowering young women in STEM. Introducing high schoolers to data science and ML concepts through hands-on projects and mentorship.
Jan 2026 - Present

Bridging Minds & Machines

Conference
📍 SPSP 2024, San Diego

My research sits at the intersection of machine learning and behavioral science — using computational methods to uncover patterns in psychological data that traditional analysis often misses.

I've applied ML techniques to psychology datasets, presented findings at academic conferences, and explored how quantitative rigor can deepen our understanding of human behavior.

AI & ML Work

01
Knowledge Distillation & In-Context Learning in LLMs
End-to-end evaluation pipeline comparing distilled vs. base LLaMA-3.1 8B models across NLI, sentiment analysis, and QA. Designed robustness metrics revealing accuracy-vs-noise tradeoffs in distilled models.
PyTorchLLaMA-3.1Hugging FaceNLP
02
Multi-Agent Procurement Analytics (UN)
Modular LangGraph system with 4 specialized agents for automated data cleaning, EDA, and financial anomaly detection across $3M+ in procurement data from 12 countries.
LangGraphIsolation ForestDBSCANMulti-Agent AI
03
Supplier Intelligence Platform (Amazon)
Production ETL and XGBoost churn prediction system that scaled platform capacity 3× and drove significant gains in data velocity, precision, and engineering efficiency.
XGBoostAWSETLPython
Get In Touch

Let's Build Something Together

Open to full-time roles starting May 2026 — Data Science, AI Engineering, and ML.