
Cody Sims
I design and build useful AI-enabled products, from accessible media tools to full-stack platforms.

Good morning!
Starting the day with code and coffee
I turn ambiguous ideas into dependable products.
Software Engineer at Microsoft
I'm a full-stack engineer who likes the space between product thinking and implementation: defining the useful version of an idea, then building the systems that make it real.
At Microsoft, I build intelligent features for an enterprise platform. Outside work, I explore AI, accessibility, and tools for everyday life—usually with a trip, recipe, or new question nearby.
Proof in the Work
Faster AI-Assisted Tasks
Copilot Buttons cut 8-minute workflows to 2 at MicrosoftMolecules Trained On
ZINC dataset behind a TensorFlow molecular autoencoderGPA at Brown
B.S. Computer Science, class of 2024Road Segments
Queried by a custom pathfinding appStudents Taught
Course materials and weekly debugging sessions at BrownCountries Explored
Curiosity beyond the codeBrown University
B.S. Computer Science - 4.0 GPA
Graduated with a perfect GPA in computer scienceBrown University
Software Engineering Teaching Assistant
Mentored 15+ project teams and built materials for a 150+ student courseCisco Systems
Cloud Networking Intern
Built REST APIs and a Grafana cost-metrics dashboardMicrosoft
Software Engineering Intern - AutoNet Platform
Cut deployment creation time by 10% with a redesigned deployments pageMicrosoft
Software Engineer - AutoNet Platform
Multi-agent AI orchestration, MCP servers, and microfrontend tooling for network engineersView DetailsWhat's next? 🚀
Selected Work

cody-sims.com
AI-Powered Portfolio & Travel Platform
Create one useful home for a portfolio, travel journal, recipes, and long-form content without turning each feature into a disconnected product.
Outcome: A production platform that supports four content domains, nine documented projects, and a reusable admin workflow.

Recipe Collection
AI-Powered Recipe Manager & Generator
Recipe sites often bury the useful content, while manually rewriting recipes makes a personal collection difficult to maintain.
Outcome: Reduced recipe capture from a manual transcription task to a guided import and review flow.

Video Description Generator
AI-Powered Accessibility for the Visually Impaired
Most video lacks audio description, excluding blind and low-vision viewers from visual context that sighted audiences receive automatically.
Outcome: Produced a working Llama 4 Hackathon prototype that turns uploaded video into timed, customizable audio description.
Rhode Island Maps
Interactive Pathfinding Application
A full-stack web application similar to Google Maps, built specifically for Rhode Island. Connected a React.js frontend with a Java backend server to create an interactive mapping experience capable of pathfinding using Dijkstra's algorithm. Utilizes a SQL database containing over 1 million road segments to calculate optimal routes between any two points in the state. Features include real-time route visualization, turn-by-turn directions, and efficient database queries for fast response times.
Molecular Variational Autoencoder
Deep Learning for Drug Discovery
Implemented a Tensorflow version of the Molecular VAE paper for automated molecular design. Trained on a dataset of 250,000 molecular graphs from the ZINC database, this deep learning variational autoencoder learns to encode molecular structures into a continuous latent space, enabling the generation of novel drug-like molecules. The model facilitates faster drug development by allowing researchers to explore chemical space more efficiently and generate molecules with desired properties.
Vunmo
Multithreaded Payment Backend
Developed a small-scale multithreaded backend server capable of receiving and transferring imaginary funds, similar to Venmo. Built in C and C++, the server handles concurrent client connections using threading, implements secure transaction processing, and maintains consistent account balances across multiple simultaneous operations. Features include user authentication, transaction history, and thread-safe data structures for handling race conditions.

Iterative Design
UX Case Study
A comprehensive UX case study utilizing iterative design principles and extensive user research to create a Point-Of-Sale System for restaurants. The project involved conducting user interviews, creating detailed personas, developing low-fidelity wireframes that evolved into high-fidelity Figma prototypes, and performing usability testing with real users. Multiple design iterations were made based on feedback to optimize the user experience for both restaurant staff and customers.

Pokemon Battle Simulator
Turn-based Combat Game
A passion project recreating the battle system from my favorite childhood game. This full-featured Pokemon battle simulator includes all 151 original Pokemon with accurate base stats, 16 type matchups with proper effectiveness calculations, and over 700 different moves with their authentic effects. Built using object-oriented design principles in Java with a JavaFX interface, the project involved scraping Pokemon data from Bulbapedia using Python and BeautifulSoup to ensure accuracy.

Pokemon Team Builder
Interactive Team Planner
A vanilla JavaScript web application for building and analyzing Pokemon teams, showcasing DOM manipulation and state management without any frameworks. Features include searching and filtering Pokemon, adding them to your team, analyzing type coverage and weaknesses, and saving teams locally. Integrates with the PokeAPI to fetch real Pokemon data including stats, types, and sprites. Built as an exercise in understanding core web technologies before learning modern frameworks.
cody-sims.com
LiveAI-Powered Portfolio & Travel Platform
Create one useful home for a portfolio, travel journal, recipes, and long-form content without turning each feature into a disconnected product.
A production platform that supports four content domains, nine documented projects, and a reusable admin workflow.
Recipe Collection
LiveAI-Powered Recipe Manager & Generator
Recipe sites often bury the useful content, while manually rewriting recipes makes a personal collection difficult to maintain.
Reduced recipe capture from a manual transcription task to a guided import and review flow.
Video Description Generator
AI-Powered Accessibility for the Visually Impaired
Most video lacks audio description, excluding blind and low-vision viewers from visual context that sighted audiences receive automatically.
Produced a working Llama 4 Hackathon prototype that turns uploaded video into timed, customizable audio description.
Rhode Island Maps
Interactive Pathfinding Application
A full-stack web application similar to Google Maps, built specifically for Rhode Island. Connected a React.js frontend with a Java backend server to create an interactive mapping experience capable of pathfinding using Dijkstra's algorithm. Utilizes a SQL database containing over 1 million road segments to calculate optimal routes between any two points in the state. Features include real-time route visualization, turn-by-turn directions, and efficient database queries for fast response times.
Molecular Variational Autoencoder
Deep Learning for Drug Discovery
Implemented a Tensorflow version of the Molecular VAE paper for automated molecular design. Trained on a dataset of 250,000 molecular graphs from the ZINC database, this deep learning variational autoencoder learns to encode molecular structures into a continuous latent space, enabling the generation of novel drug-like molecules. The model facilitates faster drug development by allowing researchers to explore chemical space more efficiently and generate molecules with desired properties.
Vunmo
Multithreaded Payment Backend
Developed a small-scale multithreaded backend server capable of receiving and transferring imaginary funds, similar to Venmo. Built in C and C++, the server handles concurrent client connections using threading, implements secure transaction processing, and maintains consistent account balances across multiple simultaneous operations. Features include user authentication, transaction history, and thread-safe data structures for handling race conditions.
A comprehensive UX case study utilizing iterative design principles and extensive user research to create a Point-Of-Sale System for restaurants. The project involved conducting user interviews, creating detailed personas, developing low-fidelity wireframes that evolved into high-fidelity Figma prototypes, and performing usability testing with real users. Multiple design iterations were made based on feedback to optimize the user experience for both restaurant staff and customers.
Pokemon Battle Simulator
Turn-based Combat Game
A passion project recreating the battle system from my favorite childhood game. This full-featured Pokemon battle simulator includes all 151 original Pokemon with accurate base stats, 16 type matchups with proper effectiveness calculations, and over 700 different moves with their authentic effects. Built using object-oriented design principles in Java with a JavaFX interface, the project involved scraping Pokemon data from Bulbapedia using Python and BeautifulSoup to ensure accuracy.
Pokemon Team Builder
LiveInteractive Team Planner
A vanilla JavaScript web application for building and analyzing Pokemon teams, showcasing DOM manipulation and state management without any frameworks. Features include searching and filtering Pokemon, adding them to your team, analyzing type coverage and weaknesses, and saving teams locally. Integrates with the PokeAPI to fetch real Pokemon data including stats, types, and sprites. Built as an exercise in understanding core web technologies before learning modern frameworks.
Want to see more?
Check out my travel logs, blog posts, and other projects.