Curated projects
Drink Sense
T1 Exercise
Loinnir
Drink Sense
This app was initially created in 2015 as an Android app with some friends as part of GlassByte, and then revisited in 2020 as a pandemic-era pet project to push as far as possible during lockdown.
The problematic loop was as follows: Consumes drinks > Delay in absorption > Drinks more > Queued drinks stack up > Too drunk. The goal of this project was to provide a neutral, science-backed early-warning observer to combat over-drinking as a result of the latency between consuming a drink and feeling the effects of alcohol in the bloodstream.
The app helps users to track and manage their alcohol consumption while providing real-time predictions of future blood alcohol content based on their intake. Users can log each drink, review their past consumption, view insights, and set notifications to monitor habits. The app encourages more informed decisions about drinking by offering a clear picture of the user’s alcohol intake and its potential effects.
Initially, the system's blood alcohol content prediction model was based on the Widmark formula, which estimates blood alcohol content by taking the user’s height, weight, and sex into account. However, later iterations evolved to use a pharmacokinetic diffusion system (the Gelabert model). It offers greater accuracy in predicting the diffusion of ethanol into the bloodstream and accounts for a wider range of physiological factors, making the blood alcohol content predictions more reliable for users, including differentiating between shots, downing drinks, and various sipping speeds over a duration.
TypeScript
Kotlin
React
Material UI
Firebase
GraphQL
BullMQ
NestJS
NodeJS
PostgreSQL
Redis
NX
Android
