441–450 of 1001 posts
Did it also to learn rust, which made it easy to convert the engine code into webassembly and create a hand score calculator for the game :)! https://cerpins.com/mahjong-tool
Plans now are to finish the client, then freshen up the calculator with some QOL and visuals, surely a bug fix might be needed here and there as well.
Labeled data is the bottleneck, so my work focuses on getting good results with less data. Key parts:
- Created EDANSA [1], the first public dataset of its kind from these areas, using a improved active learning method (ensemble disagreement) to efficiently find rare sounds.
- Explored other low-resource ML: transfer learning, data valuation (using Shapley values), cross-modal learning (using satellite weather data to train audio models), and testing the reasoning abilities of MLLMs on audio (spoiler: they struggle!).
Happy to discuss any part!
Have been building drawDB[1] for a while now. It's a database schema visualizer. Currently working on adding support for oracle databases. Wrote a parser[2] to allow importing from oracle sql. Have been struggling with motivation though, the pieces are pretty much there but I've been procrastinating on putting it all together. This has been my main side project for almost 2 years now.. I miss the feeling of novelty..…
It lets users merge two or more workouts into a single one. There have been times when I have been out riding, hiking or whatever and accidentally end the activity on my apple watch instead of pausing it. Starting a new workout means having your stats split across the two workouts.
The "usual" way to merge such workouts is to export all of them to individual FIT files, then use a tool like fitfiletools.com to merge the individual FIT files. You then have a merged FIT file, which is difficult to import back into Apple Health. This process also requires access to the internet, which is not always guaranteed when out in remote areas.
MergeFit makes this process easy by merging workouts right on device and without the need to deal with FIT files at all. It reads data directly from Apple Health and writes the merged data back to Apple Health.
The app reached a small milestone a few days ago - crossing 1000$ in total sales.
Check it out: https://github.com/amlel-el-mahrouss (it's on the pinned repositories)
Will continue to refine and possibly do more - love iterating and polishing as a way to learn.
I'm working on pure.md[1], which lets your scripts, APIs, apps, agents, etc reliably access web content in markdown format. Simply prefix any URL with `pure.md/` and you get the unblocked markdown content of that webpage. It avoids bot detection and renders JavaScript-heavy websites, and can convert HTML, PDFs, images, and more into pure markdown. pure.md acts as a global caching layer between LLMs and web content. I…
Exporting book highlights from a Kobo was slow and inconvenient; you’d need to connect the device to your computer via USB and run a script, or upload the onboard sqlite database to a website to extract book highlights. With Highlights.Email, you tap a button on-device and in a few seconds have a nicely formatted email with all your book highlights!
So, I’m just scratching my own itch mainly while learning how to build and launch something to the world. There’s two parts to this service: Rust program that runs on-device and a SveltKit app (w/ a Supabase backend) for auth and sending emails.