OLIVIA QI


About
Contact


* = concept

WEB / UX

  1. Alzheimer’s San Diego
  2. Get Sh*t Done
  3. Rate My Plate
  4. Mouseyverse
  5. Da Capo*


GRAPHICS

  1. Hasitha Guhan
  2. TeenTix merch
  3. Chinoiserie Script
  4. Brat poster*
  5. Blood poster*
  6. Luna Li merch*
  7. The Marías merch*
  8. Wild book cover*


    MISC.

    1. Surveillance
    2. My Kid Could Play That*
    3. Aphrodite Metropolis campaign*
    Get Sh*t Done    UX design, UX research, Web design  TEAM: 1 front end, 3 back end
    ROLE: Front end, mostly
    TOOLS: React Native, Python       I created this fun project for LA Hacks 2025 in 36 hours! Get Sh*t Done is a habit-tracking iPhone app that spam texts, spam calls, steals money, and tweets nonsense from your X account if you don’t complete a task.

    Visit the Devpost page  
    Github repo



    INSPIRATION
    We started brainstorming by thinking of tedious sh*t we didn't like getting in our everyday lives, and we realized there are a lot of them-- taking out the trash, doing laundry, etc. But rather than defaulting to AI to offload the work, we decided to use AI to motivate us to keep up these habits that are such an important part of being human. Of course, we had to give it a twist-- instead of being a regular habit tracker to be ignored, we made it as funny and annoying as possible so users would keep the promises they'd made to themselves.

    WHAT IT DOES

    If you complete all your self-assigned habits on time, Get Sh*t Done is just another habit tracking app. But if you miss a habit, get ready for a bunch of spam texts. Keep putting it off, and you'll be constantly declining calls from an angry AI. If you're still procrastinating, you'll probably stop when our app withdraws money from your Stripe account and donates it to your least favorite political party. Even if that doesn't faze you, you'll definitely get sh*t done when our AI tweets insane nonsense from your X account for all your followers to laugh at. Don't even think about faking it-- we need you to upload a photo of your completed task that we'll verify with image recognition for proof.

    HOW WE BUILT IT  

     We used React Native for the front end, Mongo DB for the database, Gemini and OpenAi as AI models, as well as Python, asi-1, fetch-ai, Twitter, auth(), AI-agent, React Native, Fast API, Twilio, and Expo-io. I worked with React Native for the front-end and Python for the Stripe integration 

    CHALLENGES  

     It took a while for me to figure out the Stripe integration, which makes sense since they're concerned about user privacy. By working across the front and back end, reading documentations, and following tutorials, me and another team member gathered users' payment information without initially charging them. Image recognition for verifying tasks also proved to be difficult, but we eventually trained the dataset to be more specific with what images matched up with what tasks.

    ACCOMPLISHMENTS  

     We're proud that we got so much done in such a short amount of time. We knew our project would have a lot of moving parts when we came up with it, but we didn't shy away from incorprating AI agents, LLMs, image recognition into the project because we really wanted to see the result. Besides feeling fulfilled each time our code worked, we laughed every time the AI voice berated us from the phone, and every time a silly AI-generated tweet got posted on one of our Twitter accounts.

    WHAT WE LEARNED

    Our learning experience started with coming up with a compelling product idea and we also learned about connecting front end to back end.

    BUILT WITH

    We want to smooth out the Stripe and Twitter integrations so users can enjoy a streamlined setup process, and we want to improve our image recognition model to be more specific about matching up photos with completed tasks.
    © 2025oliviarxqi@gmail.comLast updated 2/28/2025