CarlosLorenzo

CarlosLorenzo

0-day streak
FORMUFLASH I made a flashcard app aimed at spanish-speaking STEM students supporting LaTeX and Markdown (also images) syntax allowing complete control over the appearance and content of flashcards. The content is writen in Spanish as I aim to promote it and gain a userbase and the Spanish market was nonexistent (sorry english speakers, you will need google translate). Made with a django-rest backend with postgreSQL, react ts frontend and nginx as a reverse proxy. It (django and nginx )has been deployed to as individual serverless containers to gcp cloud run, it is then managed by a load balancer. This was started before I joined arcade (it has taken waaaay more hours than hours im submiting) You can access it through formuflash.com, formuflash.com(if I don't have SSL setup) (or formuflash-frontend-qkflfm2o5q-no.a.run.app, if the first link doesn't work, I don't have the domain setup yet,). To save on costs it is deployed as on demand so it might take 5-10 seconds for it to boot up. You can check the code out at: github.com/carlos-lorenzo/formuflash
https://scrapbook-into-the-redwoods.s3.amazonaws.com/0b09ff6a-e0ed-40a9-90ad-ee2f32d6b057-image.pnghttps://imgutil.s3.us-east-2.amazonaws.com/d73a4e57d8fda2a944b29b424d12c02ffdf8b77f589ec99b4fdb4465c1633289/024dc81c-0014-4548-8ad9-e3ab7003bb78.png
google emoji
goose-honk-technologist emoji
github emoji
summer-of-making emoji
react emoji
typescript emoji
I have finished with non-deep Reinforcement lerning. Implementing simple Bellman equations, dynamic programming and Montecarlo method to solve Mazes. Also, implemented q-learning to solve blackjack (41.5% achieved winrate - theoretical maximum). Maze solution with montecarlo: Github repo blackjack-qlearning: github.com/carlos-lorenzo/blackjack-qlearning
https://imgutil.s3.us-east-2.amazonaws.com/a699ec10b0ca54f29ed81c1913c76db2aed5ae6683998bef670ca02a65b00e44/39cbd5ee-ac77-46f5-a94d-796429020669.png