Hi, I'm Amelia — a Computer Science student (with minors in Data Science and English) who loves creating, writing, and untangling messy problems. I build backends and AI agents, mentor other students into tech, and spend a lot of weekends at hackathons.
›Reduced backend query latency by 60% by moving scan-based queries to PK/SK schemas and Global Secondary Indexes for enterprise indexing workflows.
›Decreased DynamoDB storage footprint by 40% by re-architecting ingestion pipelines and consolidating denormalized records across distributed microservices.
›Eliminated 10+ hours/week of manual debugging with CloudWatch telemetry and CI/CD post-deploy health checks.
›Architected and migrated highly scalable DynamoDB-backed services, driving cross-team alignment through technical documentation.
›Low-latency agent orchestration framework built for the Cornell University AI Hackathon that manages async data streams to place live outbound calls for restaurant reservations.
›Deterministic 6-state conversational state machine with NLU sentiment classification to handle rejections, renegotiations and edge cases in production.
›Concurrent Deepgram STT and ElevenLabs TTS streaming through FastAPI, driving turn-to-turn latency under 2 seconds.
A personal grindlog dashboard that tracks your coding practice — streaks, topic coverage, and recent solves — so consistent prep turns into visible progress.
›Led dataset discovery and model building, sourcing the Heart Disease Health Indicators dataset (Kaggle) and PhysioNet activity/sleep data to train a risk classifier.
›Collaborated across a 3-person team on data pre-processing, model analysis and results visualization for a client-facing deliverable.
›Grounded feature selection and interpretation in WHO cardiovascular disease guidance to keep the model clinically meaningful.
A Python + Flask toolkit that compares genomes and proteomes, computing Average Nucleotide Identity (ANI) and Average Amino Acid Identity (AAI) and rendering identity heatmaps.
›Ships as both CLI scripts and an upload-and-visualize Flask web app, outputting PNG heatmaps and Excel workbooks.
›Wraps pyANI-plus (fastani, anib, anim, dnadiff, sourmash) for ANI, and implements a phage-compatible AAI pipeline with a k-mer prefilter and bidirectional best hits.
›Runs fully in Python with no external alignment tool requirement for AAI, making reproducible bioinformatics analysis accessible locally.
›Serverless developer utility built for the Google Gemini AI Hackathon that recursively walks the GitHub API to map unfamiliar repositories into architectural roadmaps.
›Data parsing pipelines over complex file structures and syntax trees, wired to AWS Bedrock search to output reliable docs that accelerate onboarding.
›Designed 20+ dynamic UI components with React Native and Tailwind CSS for a friendly, guided interface.
A hardware-plus-software system for infection control in hospitals, ranked top 10 out of 200+ projects.
Tech stack
JavaSpring BootMySQLSQL
Highlights
›Real-time sensor ingestion service normalizing high-frequency CFU telemetry into a structured event stream persisted to MySQL.
›10+ table relational model supporting analytics queries and event-driven alerting.
›Achieved a 95% improvement in the efficiency of the communication platform between devices and staff.
Where2Meet
2024
GIS-powered smart meeting point finder
Tell it where everyone is coming from and it finds the fairest place to meet, with directions, weather and travel times.
Tech stack
GISGoogle Maps APIPythonAI/MLWeather APILangChain
Highlights
›Built core functionality for providing directions using GIS and the Google Maps API to locate optimal group meeting points.
›Implemented AI/ML algorithms to analyze location requests and determine the most efficient meeting points for groups.
›Integrated the Weather API and LangChain API for real-time weather data and travel time estimates.
TV Recommendations System
2024
AI-powered viewing prediction engine
A recommendation engine that learns viewing patterns at scale and suggests what to watch next.
Tech stack
PythonDeep LearningMachine LearningRecommendation Systems
Highlights
›Coordinated with cross-functional teams to integrate an AI-powered recommendation engine, resulting in a 40% increase in user retention and satisfaction.
›Designed and implemented ML models — including deep learning techniques — to analyze large-scale viewing patterns and predict user preferences with high accuracy.
Image-to-mood playlist generator · Django + Google Vision + Spotify
Upload a photo and get a playlist back — the app reads the image's colors and mood with Google Cloud Vision, then builds a matching Spotify playlist for you.
Tech stack
PythonDjangoGoogle Cloud VisionSpotify APIHTMLCSS
Highlights
›Django web app that analyzes an uploaded image's dominant colors and maps them to emotions, then assembles a Spotify playlist matching the detected mood.
›Integrates Google Cloud Vision for image understanding and the Spotify Web API for track retrieval and playlist generation.
›Persists search history and mood mappings in a Django data model so recommendations stay reproducible across sessions.
I major in Computer Science and I have a passion for Technology, Innovation & Business. In the future, I hope to merge my passions and my enjoyment into my career — which is why I'm exploring Software Engineering and Product Management roles.
Gettysburg College
B.S. Computer Science · Minors in Data Science & English
Sep 2022 — May 2026 · Seattle, WA
Gold Medal, EUROINVENT
Involvements
Google Developer Student Club — President
NYU Girls Who Code — Mentor
CodePath — Community Member & Student
Rewriting the Code
Society of Women Engineers
05
Languages & Certifications
Languages
EnglishNative / Bilingual
VietnameseNative / Bilingual
FrenchProfessional working
SpanishElementary
Certifications
Introduction to Career Skills in Software Development
Generative AI for Product Managers
Foundations of Geographic Information Systems (GIS)
Succeeding in Web Development: Full Stack and Front End