TONY WANG
> studying cs + math @ stanford
projects
a compilation of things i've built!
Shepherd
1st Place Grand Prize ($12,000) at TreeHacks 2026. A smart cane that guides indoor and outdoor navigation for blind people, costing ~1/20 of existing devices and providing better guidance.
Mercury
HackMIT 2025 1st Place Fetch AI Prize and Final Panel Presenter. Built a smart home sustainability AI agent that automatically adjusts lighting, temperature, and more based on climate conditions.
Aegis
1st Place at xAI x Vercel Hackathon. Built autonomous emergency response agent that can process IDs scans, maintain real-time location tracking, and conduct calls for emergency guidance.
ContractMe
2nd Place at Brex AgentJam. Built AI agent that automatically handles essential services contracting, from a voice-based phone agent handling requests to a full-fledged contract management system.
Code Canary
1st Place Overall + 2nd Place Modal Prize at the AI Agent & Infra Hackathon by Lux Capital. Built AI agent that generates SBOMs, detects vulnerabilities, and triages transitive dependency risks.
Sonus
2nd Place Coinbase Prize at YC Agentic Payments Hackathon. Built automatic payment verification system for consumers with Locus.
A LIVING POEM
Infinitely self-rewriting poem insprired by A LIVING POEM by sasha stiles, an art piece at NYC MoMA.
Constellation
An engineering-first recruiting system that visualizes talent networks as a 3D galaxy. Calibration shows precision@20 with gold standard markers.
EXPLORAGON City Exploration Game
City exploration game that divides SF into 1,000 H3 hexagons. Players complete AI-verified, location-specific challenges to claim territory in real time, with live leaderboards and multiplayer GPS tracking. Top 25 at FAF Multimodal Hackathon.
Small Claims Court Form Generator
With Korbin and Benny, built a Next.js tool that turns a simple web form into a fully filled California SC-100 PDF. Features AI-assisted legal text generation, court lookup, business info autofill, and smart validation.
Caply (User Data Management Tool)
Akaash and I built a tool to streamline California Consumer Privacy Act (CCPA) compliance for SMBs. Caply empowers companies to efficiently identify and purge personal data from their systems, as well as automatically notifying third-party vendors, ensuring adherence to critical privacy regulations.
AEMP Web Scrapers (Stanford Public Interest Technology Lab)
Led a web-scraping project with the Anti-Eviction Mapping Project (AEMP) as part of the Stanford Public Interest Tech (PIT) Lab. This project automates the collection of critical housing data to support AEMP's mission of providing data for fighting housing insecurity.
Laboobie
We made a Labubu dance! Laboobie can FaceTime you, call your phone, swap between being an AI girlfriend or boyfriend depending on which Labubu you hold up, and even challenge you to Red Light, Green Light from Squid Game. Won 2nd Place at the AI Valley x Pebblebed Hackathon.
SEC Analysis Tool
Built a web app that allows users to upload SEC filings like 10-Ks or 10-Qs and ask questions about them. Supports consistent file storage and multiple conversation threading.
Ensemble AI
Built orchestration layer for ensemble multi-model video captioning, orchestrating Qwen, Claude, Nova, and YOLO, achieving quality gains in anonymous human eval tests and hallucination F1 scores.
Monet (Voice-based Tool for Google Slide Editing)
Developed a voice-controlled tool for seamless presentation editing for the OpenAI x Pear VC hackathon; leverages voice commands to manipulate slide content, making accessible the editing workflow.
Neurosense (DSM-5 & SAMHSA-informed Crisis Care)
Coded 'Neurosense,' an expert support and crisis care platform informed by DSM-5 and SAMHSA guidelines. This application provides real-time assistance, bridging the gap between individuals in need and professionally-backed mental health advice/resources.
Hourglass (Productivity & Time Management iOS App)
Developed 'Hourglass,' a productivity-tracking iOS application designed to gamify time management. The app fosters motivation through social accountability features and provides visual progress tracking, empowering users to take control of their time.
MultiSport (AI-Powered Sports Analytics)
Designed and implemented real-time CV system for simultaneously live tracking multiple players in sports footage, with bird's-eye view analytics for insights in player movement and strategy. Implemented Kalman filtering, homography transformation, Canny edge detection, etc.
ACM ML Lab (Avian Species Classification)
As part of the ACM ML Lab, developed 'Bird-ResNet,' which applied deep ResNets to the task of avian species classification from images.
CS44N (Computational Graphics)
Completed a series of projects in computational graphics mentored by Professor Doug James, exploring topics from ray tracing to 3D modeling. Learned a lot about the math behind modern computer graphics.
CharacterGraph (Narrative Analysis Tool)
Got confused keeping track of the characters in Cixin Liu's novels, so I made 'CharacterGraph,' a tool that analyzes a body of text to automatically generate and visualize a graph of character relationships. This provides a super fun at-a-glance graph of narrative structures.
HeadcountAI (AI for Headcount Forecasting)
Designed and proposed an AI Technical Program Manager for project forecasting. I integrated LLMs and deep neural networks to create a more accurate and automated system for workforce planning. Received Amazon Raising the Bar award.
CS109 (Stanford Open Policing Project Analysis)
Conducted statistical analysis of data from the Stanford Open Policing Project. I explored patterns in traffic stop data and found that black and hispanic drivers are more likely to be searched when pulled over than white drivers. Awarded 3rd place in the Stanford Probability Contest.
Stanford Management Group (SMG) Product Redesigns
Worked with Microsoft Azure (2024), Fizz (2024), Tesla (2025), Pear.us (2025) to redesign their product pages and landing pages. Developed design specifications and strategy recommendations through user interviews, competitive analysis, and Figma design testing. Some are now live!
Indica Rice Genomic Research
As part of the Non-Trivial Fellowship, I used computational gene expression analysis and machine learning to uncover biomarkers for drought resistance in Indica rice. Identified three key hub genes (D2HGDH, DMC1A, and COQ3) with known protein structures, offering tractable targets for improving crop resilience.
TF-CNN Bioinformatics Tool (SAIL Kundaje Lab)
Contributed to computational genomics research during an internship at Stanford's Kundaje Lab (SAIL). Focused on developing and interpreting transformer/CNN models to better understand transcription factor binding motifs.
Myocarditis Bioinformatics Research (Stanford Medicine Sean Wu Lab)
Worked at the Stanford School of Medicine in the Sean Wu Lab, contributing to a CRISPR myocarditis study. Gained hands-on experience in dimensionality reduction and clustering.
