Arjun Mann

Software Engineer Intern @ Lumber Jul 2025 – Current

• Cut customer onboarding costs by building an automated pipeline that scrapes PDFs (PyMuPDF, Tabula API) for payroll information and loads the results into folders using Google Drive API.
• Developing an internal full-stack dashboard using React, Node.js/Express, and MySQL, deployed on Azure, to streamline the configuration, monitoring, and validation of payroll data ingestion, with job orchestration via CRON and cloud storage in Azure Blob Storage.

Software Engineer Intern @ Metric Analytics Apr 2025 – Jul 2025

• Developed an in-house standard RAG (Retrieval Augmented Generation) model for scoring AI generated image content using HuggingFace transformers, replacing expensive API calls and cutting inference spend by about 60%.
• Reduced average model response time from 7 to 4 seconds through query parallelization, improving data security and increasing videos scored per shift by about 30%.
• Ran controlled environment experiments to understand tradeoffs across 10+ parameters (ex: temperature, top-p, model sizes, etc.), establishing baseline metrics for the new model I created.

Research Assistant @ UCI Mar 2025 – Jun 2025

• Collaborated with Professor Christine King and PhD candidate Dalton Salvo on the development of a volumetric video based VR platform for clinical immersion in medical training
• Automated 3,000+ OBJ/MTL/texture imports via Python scripting and metadata parsing to batch import assets into UE5 while renaming materials to prevent overwrite conflicts
• Optimized Unreal Engine imports for VR interactivity by collaborating with the research team to integrate assets into fully navigable VR spaces, reducing rendering bottlenecks and enabling scalable processing for future volumetric capture sessions.

Graduate Student Research Assistant @ UCI Sep 2024 – Mar 2025

• Collaborated with graduate student Samyak Jhaveri supporting a thesis on quantum computing and GPU accelerated software engineering.
• Scraped 12+ GitHub repositories of C++ programs, extracting pragma for loop pairs and organizing them by clause type (collapse, reduction, deviceptr) for LLM fine tuning
• Trained an LLM on 200+ files, optimizing CPU serial programs to be up to 20x faster.

Software Engineer Intern @ Lumber Jul 2024 – Sep 2024

• Developed a scraping application to streamline 300+ projects for an LLM, using Selenium to automatically log and return JSON data for account information, fully automating a manual process.
• Generated timesheet data for 1000+ employees for a CRM, using Playwright and Beautiful Soup to convert internal data (i.e. time, employee information) into a universal CSV format.
• Cut customer onboarding costs by building an automated pipeline that scrapes PDFs (PyMuPDF, Tabula API) for payroll information and loads the results into folders using Google Drive API.

Events Committee Member @ ICSSC Jul 2024 – Sep 2024

• Had fun planning and volunteering for events to engage UCI ICS students and provide them a mix of fun activities and career resources.
• Helped organize IrvineHacks 2025 by hosting a MongoDB workshop to engage up to 400 participants.
• Facilitated panels with Computer Science faculty and alumni to connect 40+ attendees.

3x Mentor @ UCI Hackathons Sep 2024 – Current

I've mentored at ZotHacks 2024, Soar Datathon 2024, and VenusHacks 2025. Mentoring at these large events requires applying and being present the whole event. For ZotHacks 2024, I was tasked to take ownership of one team to mentor. For the other two hackathons, I went around helping any team that requested it. I do it to because I love teaching and helping others build.

Python Tutor @ Hack Club Dec 2023 – Mar 2023

I volunteered as an intermediate and advanced level instructor at Winter Camp for Hack Club to teach middle schoolers how to code. I created teaching materials from scratch with intuitive examples, slides, and projects for students to follow over 2 week courses. For the intermediate course I taught data structures and algorithms, and for the advanced course I went in depth with fun libraries like Pygame and Pandas and taught their uses with real-life examples.

Goober Search

‹ ›

Goober is a web search engine I collaborated on with 4 others. We got it to handle tens of thousands of pages while maintaining a query response time under 300ms! We tokenized both queries and documents with snowball stemming, and used an inverted index with tf-idf scoring to rank and display the top results. This was my first time working with a scaled dataset. To accommodate 40k+ files, we preprocessed junk sites and used map reduce techniques to prevent overloading our puny laptops' RAM.

Fablix Website

‹ ›

Fablix is one of my most ambitious web apps. I made it with my roommate. It's a movie catalogue site with features like advanced browsing, cart checkout backed by sessions, full-text search with autocomplete, bot checks with reCAPTCHA, and more! We ran it on the cloud with AWS EC2 instances using Apache Tomcat. We created our own ETL pipeline for xml files using a DOM parser, and the app uses MySQL connection pooling. We even experimented with containerization by creating a docker file and running it in on Kubernetes clusters (1 admin, 2 workers). This really accelerated my interest in cloud computing and I look forward to expanding my skills here! :)

Zotify

‹ ›

At Zothacks 2024, I mentored a group of four novice CS students and showed them how to build a web app from scratch. I taught them Flask, React with Vite tooling, and git. I also divided them into teams, debugged their code, and gave them more advice than they probably wanted about coding web apps. Together we built a song recommendation site in 48 hours that calls Rapid API in a Flask backend to pick from 100 million songs based on a playlist the user creates using search.

Echo Words

‹ ›

• Created an interactive app to translate uploadable manga and foreign text.
• Integrated the Gemini Developer API with prompt engineering in a Flask backend to improve speed and accuracy.
• Tokenizes sentences to provide highlightable word or phrase translations.

Red Goes Fasta

‹ ›

• Received “Fastest Game” award at VDGC Winter 2024 by creating a Mario-Kart style racing game in Unity
• Implemented the driving, collision effects, and powerups by creating their physics mechanics through C# scripting
• Collaborated with programmers, artists, voice actors, and game testers through weekly meetings

Afterword

These are just some of the cool projects I’ve worked on. I selected these ones because they each showcase something significant I learned that shaped how I approach programming and collaboration. Feel free to ask about my other works :)