I worked primarily on AI Dungeon -- Latitude's flagship product. AI Dungeon is one of the largest AI consumer applications in the world. >1M monthly active users. Learned a ton about language models, serving them in production. Unit economics etc etc. Became a huge Mixpanel nerd.• Lead two major revenue projects in Q2-Q3: Free Tier Monetization and AI Dungeon on Steam• Monetization - Shipped AI Dungeon’s mobile and web advertising platforms, which generates revenues of ~$75k yearly. Aligned ad strategy with unit economics of player usage and compute costs.• Steam - Launched a successful wishlist campaign, generating 45,235 organic list-ads. Grossed $100,000 in two weeks after launch. Managed delivery of copywriting, marketing assets, trailer videos, electron builds and conversion attribution tracking. Became featured game on Steam Store.• Q3: Authored AI Dungeon’s testing infrastructure: continuous running multi-platform and browser E2E tests for every screen on the app.• Q4: Built an extensive Moderation Platform to enable moderation of hundreds of content and users every day.• Q4: Researched and led implementation of Latitude’s events architecture on Mixpanel for superior analytics insights and scalability. System processes over 500k events per day.

• Worked with Dr. Porter Jenkins• Implemented PIXOR on internal dataset, which involved adjusting the model to work on data outside of the realm of self-driving. Results were contributions to academic paper submitted to NeurIPS for state of the art 3D object detection in highly occluded scenes• Contributed to orchestration scripts that converts LiDAR pointclouds into aerial view images used for 2D detection

We were trying to create autonomous companies. Starting in e-commerce. A full AI-enabled pipeline from product ideation, manufacturing and customer acquisition. Worked with huge music labels. I learned about Machine Learning for the first time. Spent most of my time building mvps, traveling to see customers. • Lead product and all non-ML development on Genesis, one of three core company teams. Generally touch every aspect of work, from setting product roadmap, doing customer sales, product design, internal/ external tooling.• Created contractor management infrastructure from scratch to support scaling from 0 to 12,000 projects/ year pace over 40 days. System also automates operations: NDA signing, notifications, project specification, time tracking.• Trained a finetune object detection model with YOLOv5 to detect logos on merchandise. Collected and labeled data personally. Served as the foundation for post-processing pipeline that was built after.• Built user role management into the app for internal and customer use. Created the customer admin dashboard for user role management into the Web App with React, Postgres and Koa. Used to manage thousands of accounts.

• Commissioned to design and build a ML deployment pipeline as part of University Capstone Project.• Designed architecture to replace existing lambda-based pipeline that provides the following: decouples lambdas from ML computing, removes model filesize limits, versioning system for models, containerizing, one-click frontend deployment • Currently working to create a complete deployment solution based on MLFlow, AWS, and React

Ship 1 -> Integrated with UI notification service for exports• UX improvement for one of the most used features (exporting survey results, millions of requests)• users used to wait to download exports, now they get an asynchronous notification• integrated our export service with Qualtrics' internal queue• huge improvement to the export experience (polling -> async notifications)Ship 2 -> Reduce startup time for a critical security service by >90% (15 minutes -> 1.5 minutes)• central webservice for facilitating encryption and decryption company-wide (100ks / request per day)• a combination of redesigning storage schema + implementing additional cache layers for encryption keys reduced latency, and lengthy (and growing) start time. Led to lighting fast deployments and quicker incident resolution.• planned and implemented cutover strategy and proposed future plan Other Ships • Wrote endpoint for historical version data for a response• Designed and proposed BYOK compatibility for v1 endpoints of a service (BYOK was previously only available in v2)

• Worked with a microservice architecture on the Storage Platform Team, which overseas all survey data imports and exports from the Qualtrics Platform. • Project focused on improving performance and reliability of the import flow -> foundational feature to Qualtrics surveys (millions of requests per week)• Reliability: implemented a major milestone feature that allows imports to be interrupted and resumed.• Leveraged Redis, S3 to manage and retrieve complex job state• Resume feature enables the following: eventual removal of upload file size limits, flexible deployment (lengthy imports will auto resume after host restart), seamless recovery (import picks up where it left off even if server crashes) • Performance: reduced latency by deprecating a legacy php microservice by rewriting its features into another service (java / scala) • Proposed design of new internal endpoint in a golang service that enables "upsert" functionality

• Prepared entire app for v1 release. Documentation, Black Box Testing and migration from Gatsby.js to Next.js• Built custom auth solution that integrates Magic.link and Airtable to create a password-less, role-based system that can be easily managed via the Airtable interface.• Optimized state and drastically reduced client-side http request volume by over 50% using react-query.

• First technical employee. Remote work.• Facilitated major platform transition away from third-party SaaS software to in-house software by rebuilding features from scratch. UI design with Sketch. • Responsible for expanding CI processes and staying current with new offerings. Integrated Travis CI.• Responsible for following user touch points: online webform, kiosk interface, promo redemption screen.• Frontend written on gatsby.js, tests written with react-testing-library.• Endpoints written in express, backend tests with jest.

• Remote work as designer and frontend developer• Redesigned and developed coupon redemption flow• Created component library with Emotion.js and React, pages generated through Gatsby.js

• Developed and deployed production code on the BYU App (~60,000 active users) in Swift. • Built client-side caching into the app. Built on UserDefaults with configurable intervals and manual overrides.• Upgraded timesheet feature to allow complex active and retroactive work-order edits, five new view additions, and integrations to more than five distinct web services. • Set up automated Runscope tests.• Owned Off-Campus Housing feature upgrade project. Overhauled UI and integrated with new web service.

• Hired as private consultant, remote work. Delivered project successfully within two week deadline.• Discussed UI and app functionality needs with emphasis on a “premium” feel and iOS-like interface.• Prototyped wireframes in Sketch, developed with Swift.• App featured on KickStarter page (launched September), 80% of $20,000 goal funded within first week. Project eventually 292% funded.

• Managed two adwords campaigns for two 501(c)-classified theatres with a monthly budget of $10,000.• Streamlined and enhanced A/B Testing, task automation and automatic report generation through various software such as Hotjar, Zapier, Google Data Studios, Google Analytics, Tag Manager, and more.• Completed and received Google Adwords Certification.

Received education about the hedge fund industry, which includes tax-beneficial and investor-specific company structures, various local regulations, and different trading strategies: long/shorting and margin trading.Updated the office’s entire physical database with self-devised filing system that maximizes efficiency of space usage via customized allocation of storage.

Received one on one training by executive managers of each of the following departments: Retail, Wholesale, Marketing, Design. Provided market research to Marketing Manager and discerned trends through various media channels.

Primary assistant to CEO, filed and organized online database, contacted over 50 prospective clients through cold calling and emails