code, memes and maths

not sure why i started this, the idea was to clear out some concepts that i am weak at. with this project i can basically sharpen 3 things, python, maths and next.js.

here is my reason to do it exactly – i cannot write python, people say it has the easiest syntax, i beg to differ c++ has the easiest one, python is so confusing to me, brackets are the most perfectly designed things after cars and video games. the whole idea to ditch that concept to have spaces is revolting to me. think about it i can barely see a computer screen how do i notice this      i put 4 spaces ~ 1 tab. it might be visible here but in code it is sort of hard. and there is no restrictions on things you can do on list like have atleast some basic ones right?!

anyway i figured i should atleast be able to write basic code in python. and should be able to manipulate data and visualize it with pandas and numpy and matplotlib that feels sorta basic.

maths – i am not bad at math but i did see a lot of limitations/knowledge gaps when i was studying ai, so i figured why not now.

next.js – not gonna lie i do not like it that much especially with tailwind looks yuck, like when angela yungmi if you don't know who she is said spaghetti code i am sure she was talking about next with tailwind but industry is moving towards it so why not. next feels like a failed attempt on creating a flutter copy. no matter how many stacks i switch flutter would always be <3

also a little break once in a while from strands can be important for my hair routine

i also like documenting stuff and am a nerd (ps this my 3rd attempt at doing this).

a developer who hasn't made his own blogging website is not a developer

this will probably be the last project before i go deep into the mindless job applying arc. gonna build this one out with a friend. simple as always: a swiftui application that can analyze music but here’s the kicker no apple music links, only spotify (not really a kicker tbh). the app will also suggest similar music based on what you just heard.

i think the only thing that is left in my arsenal apart from a vs code fork is a good swiftui app and a bit of an ml brush-up. this one should hit that spot perfectly. also the inspiration came from this video and lets face it shazam sucks for just giving out apple music links.

a good future improvement would be auto-creating a spotify playlist with the current and recommended songs. i might not release it as an actual app though probably just a homebrew package or repo. app store publishing is too much overhead, and testflight exists but no recruiter is gonna ask for a testflight link.

right now, starting out, i have 4 questions:

1. how do you even recognize music (no api calls to an llm)
2. does spotify have an sdk or api that doesn’t involve too much auth pain to get track links
3. swift is new territory for me
4. hoping this group project doesn’t end up solo (especially because strands should be wrapped by end of august and it’s already the 4th)

probably gonna start by firing up a jupyter notebook to see what i’m dealing with here.

riff - a dumb little app to recognize music, link spotify, and flex some on-device ml. not trying to change the world, just the playlist.

so many people have their own blogs on their personal websites but most of them are old posts, and the only traction those posts get is when someone randomly follows them on twitter and stumbles across their site. so i thought about this having an extension where you can track your favorite creators. i’m not sure, maybe that’s what RSS was for, but clearly i’m not uncnew slang for boomers. so i’ll never know.

the idea is simple: add your favorite site links, and if anything changes on that page, you get notified. i wanna keep this project simple and not too complicated, so no supabase, no backend running across a dozen services. no email updates just a ping on the extension that says

look, your favorite nerd made an update.

maybe in the future (or if cursor spits out some magic), i’ll add something like “x users are keeping up with your site” would be a cool touch. and obviously, if this isn’t your thing, you can always go the normie route with substack, medium, dev.to, or whatever. but if you’re reading this, you’re probably a nerd so hang on while i vibe code this listening to the weekend.

gonna call this one caterpillar 🐛 - no idea why but sounds cool

🐛 https://github.com/shresthkapoor7/caterpillar

always wanted to keep track of how many applications i'm sending out. thought about using notion but it's too bloated for me. they have notion ai but it's mid can't even edit databases properly, tried so many times, never worked. could use excel but i still need a mouse for that so yeah, ngmi.

anyways i had this local folder with my latex resume running inside cursor (pretty sure i set it up last week or so). the idea was that i could tweak my projects based on the jd. then i thought, what if i just track all my applications here too?

fast forward and i’ve got 3 shell scripts written (technically 4, but using 3), with a folder structure like:

resume/
├── main.tex
├── job_applications.json
├── create_application.sh
├── list_applications.sh
├── build_resume.sh
├── applications/
│   ├── README.md
│   └── 2024-07-28_Google_Software_Engineer/
│       ├── main.tex
│       ├── application_info.json
│       └── resume.pdf
└── README.md

1. create_application.sh when i run this, it creates a folder for the application and drops in a copy of my main.tex file. i can make edits to tailor it for the role.

./create_application.sh "google" "software engineer"

2. list_applications.sh i use two json files: job_applications.json and application_info.json. the first one tracks everything - id, company, position, date, folder name, status, all that. when i run the script below, it lists out all the applications i’ve made so far.

./list_applications.sh

3. export_applications.sh this one’s my favorite. every application folder has its own application_info.json with extras like job url, contact person, notes, etc.

./export_applications.sh

this script dumps everything into a csv with s.no, company, position, date_applied, folder, notes, status, job_posting_url. and since it’s csv, i can even run analysis on it 🤓 (yes i know, i’ve gone full nerd).

i can write markdown in the blog posts and it will be rendered as html. i can write tooltips tooltipi am a tooltip.

i can write code in the blog posts and it will be rendered as well i can run it in the browser.

PythonRun
print("hello world")
Loading Python environment...

i can also share links to individual posts

i can make tables in the blog posts.

i can also embed tweets but with a caveatimage/video quality won't be perfect and no this is not iframe