Collin Boler

Senior at Princeton studying Computer Science & Machine Learning.

Collin Boler

Experience

Oracle - Cloud Engineer Intern

Summer 2026

  • Engineered a scheduled workflow to detect and diagnose performance drift in production machine learning models, decreasing median time to resolution by over 50%.
  • Developed an agent that optimizes Spark compute clusters in Oracle AI Data Platform workspaces, improving compute cost efficiency by 30% across benchmark workloads.
  • Built an internal tool for Oracle Fusion AI Agent Studio that automatically executes scheduled agent evaluations, saving over 10 engineering hours per month.

Credit Benchmark - Data Science Intern

Summer 2025

  • Architected a data pipeline to back-test 10 years of credit data from 40+ global banks (110,000+ entities) and authored the resulting industry ↗ white paper, synthesizing technical findings for non-technical stakeholders.
  • Engineered and optimized a supervised machine learning model to forecast North American corporate credit trends. The model enables the firm to identify future sector-level credit deterioration or improvement.

1440.io - Software Engineering Intern

Summer 2024

  • Developed a customer outreach module that eliminates language barriers by integrating transcription, translation, speech synthesis, and response recommendations, which boosted client productivity by over 50%.
  • Contributed to the design, development, and maintenance of custom Salesforce applications for enterprise clients (Lowe's, Formula 1, LEGO) leveraging Jira for task management.

Selected Projects

LeetCode Wrapped ↗ leetcodewrapped.com ↗ github

Query LeetCode's GraphQL API to generate a Spotify Wrapped-style summary of your yearly coding activity.
Used by 900+ developers.

Recall AR ↗ white paper ↗ demo ↗ github

Research exploring how augmented reality glasses can support social memory through timely visual cues. RecallAR recognizes consenting enrolled people and displays their name alongside a relevant conversation prompt, helping wearers recall past interactions and naturally continue conversations.
Selected for Princeton COS IW — AR Meets AI.

Computer Vision Geo-Location Chrome Extension

Chrome extension leveraging a computer vision model trained on 5 million images to predict the precise geographic location of on-screen images.
37,000+ installs and 9,500+ weekly active users.

CRXray ↗ npmjs.com/package/crxray ↗ github

Git Clone for Chrome Extensions. NPM package that lets you clone any Chrome extension's or Edge add-on's deobfuscated source code to your local machine. Also features security scanning and beautification. Try it out with: npx crxray <extension_url>
Downloaded by 1500+ developers.

Publications

Discriminatory Power of Consensus Ratings ↗ link ↗ pdf

Credit Benchmark, November 2025. Research analyzing how Credit Benchmark's Credit Consensus Ratings compare to S&P Global Ratings in identifying default risk. Results show CCRs demonstrate strong discriminatory power over the past decade, performing on par with S&P's ratings.

Recall AR: Augmented Reality Memory Layer for Social Interactions ↗ pdf

Princeton COS IW 04: AR Meets AI – Augmenting Humans in the Physical World, May 2026. Research exploring how augmented reality glasses can support social memory through timely visual cues. RecallAR recognizes consenting enrolled people and displays their name alongside a relevant conversation prompt, helping wearers recall past interactions and naturally continue conversations.

Athletics

Princeton Varsity Track & Cross Country

3:58 mile PR, Division I All-American, Ivy League champion. Commit 30+ hours weekly to training and competition.

Collin Boler with his Princeton teammate at USAs
With my teammate at the USATF championships this past summer. To qualify, both of us ran a time that placed us among the top 36 1500m runners in the United States.

Awards

Most Innovative - Oracle Intern Hackathon

Developed an agent evaluation workflow that uses a fleet of specialized testing agents to evaluate a target agent across custom edge cases.

3rd Place - Sentient AGI Real Revenue Hackathon

Planned, developed, and launched a full-stack web application using Sentient's inference model to help small businesses save time on advertising.

The Babb Trophy

Awarded annually to the freshman member of the track team who through sportsmanship, influence, and performance has contributed most to the sport at Princeton.