Hi, I'm George

About

I'm an undergraduate student in the University of Michigan pursing a BSE in Computer Science. I'm passionate about applying technology to deliver scalable impact. I've built production systems as a founding engineer, worked in bank tech, shipped an LLM tool deployed across 26 sites, and published first-author research. Outside of code, I'm interested in philosophy, game design, bouldering, skateboarding, golf, guitar, and calisthenics.

Experience

Building and scaling production systems.

Capital One

Software Engineer Intern

Jun 2026 – Aug 2026

98%coverage vs 80% gate

• Built the PUT /certify-asset endpoint (one of 44 REST endpoints) behind a three-tier authorization model, shipping it with React badge components to production at 98% coverage against an 80% gate.
• Architected 2 AWS EventBridge-scheduled Python Lambdas: a daily certification-reminder digest querying PostgreSQL views to email owners, and a ServiceNow reconciliation pipeline uploading encrypted CMDB reports to S3.
• Authored idempotent Flyway PostgreSQL migrations building a ServiceNow CMDB caching layer (2 tables, 2 views, 2 indexes) within a 52-migration schema history, optimizing cross-table SQL for automated compliance reporting.

AWS Lambda

EventBridge

Python

PostgreSQL

React

BoilerVault

Founding Engineer · Contract

Jan 2026 – Present

$130K+revenue reconciled, 0 manual entry

• Reconciled out-of-order payment events across 4 Stripe accounts and 3 WordPress booking sites via idempotent webhooks, processing $130K+ gross revenue with 0 manual entry.
• Built a multi-tenant ops platform (FastAPI, PostgreSQL, Next.js/TypeScript) for a 3-campus storage business with JWT auth, RBAC, 20+ endpoints, and 140+ tests.
• Migrated 2,000+ legacy records via three-tier matching (exact email → fuzzy name → Stripe charge fallback), replacing a failing Zapier workflow as the new production system of record on Railway/Vercel.

FastAPI

PostgreSQL

Next.js

TypeScript

Stripe

Nexteer Automotive

Software Engineer Intern

May 2025 – Aug 2025

87.5%less code-review time

• Deployed an LLM-powered IDE extension across 26 sites, cutting code-review time by 87.5% for safety-critical embedded steering systems.
• Built the extension (Python/TypeScript, Azure AI + Copilot APIs) to parse 300+ engineering guidelines and automate compliance checks across C/H files.
• Tuned prompt pipelines and few-shot strategies to 95% violation-detection accuracy, saving engineers ~8 hours of review per week.

Python

TypeScript

Azure AI

LLM

Prompt Engineering

Projects

A mix of systems, ML, and full-stack work. Here are a few I'm proud of.

Preprocessing

0.97Jan 2026 - Apr 2026

validation AUROC

Image Classification & Transfer Learning

Built and compared a CNN and a Vision Transformer implemented from scratch in PyTorch (patch-to-token projection, a learnable [CLS] token, sinusoidal positional embeddings, multi-head self-attention) for dog-breed image classification. A 2-stage source/target transfer-learning pipeline pretrains on a 10-class breed task, then fine-tunes a frozen convolutional backbone on a binary target task, sweeping augmentation intensity across 4 levels and tuning early-stopping patience on validation AUROC to correct overfitting, reaching 0.93 accuracy and 0.97 AUROC.

PyTorch

CNN

Vision Transformer

Transfer Learning

Architecture

3,000+Jan 2026 - Apr 2026

Wikipedia docs indexed

Scalable Search Engine

A MapReduce pipeline builds a TF-IDF inverted index over 3,000+ Wikipedia documents. A Flask REST index server scores queries with PageRank-weighted cosine similarity across 3 partitioned index segments. A concurrent, service-oriented search frontend dispatches threaded requests to each segment and merges the ranked results. Segments and index servers scale independently for horizontal scalability.

Python

MapReduce

Flask

TF-IDF

PageRank

100+Jan 2025 - May 2025

users served

Resume Screener

NLP pipeline over 1,000+ resumes and job postings. scikit-learn and TensorFlow ranking models with sentence-transformer embeddings and RAG, deployed via Streamlit to 100+ users.

Python

scikit-learn

TensorFlow

RAG

Streamlit

GitHubLive Demo

Skills

Education

Research