Ronak Vusirikala

Ronak Vusirikala

B.S Business Technology & Management

Projects

Skills

Experience

Resume

Contact

Ronak Vusirikala

Ronak Vusirikala

B.S Business Technology & Management

Projects

Skills

Experience

Resume

Contact

Ronak Vusirikala

Ronak Vusirikala

B.S Business Technology & Management

Sift

Sift

Sift

Apple Store Web App

Apple Store Web App

Apple Store Web App

Two iPhones 15s displaying a Web3.0 learning app

Overview

Overview

Sift is a native macOS app I built that sits on top of your existing Apple Music library and subscription, basically a smarter client for Apple Music the same way people already use third-party apps for things like email or calendars. It’s built on Apple’s own MusicKit framework, so it talks directly to your real Apple Music library and catalog rather than faking or scraping anything. I designed and built the whole thing myself, from the interface down to the listening-history logic that actually powers the smart features, and it’s in the final stretch of App Store review right now. The core idea was pretty simple: Apple Music has a massive catalog and I already pay for it, but the actual experience of using it feels flat compared to what Spotify has trained people to expect. THis led me to I build Sift as a layer to Apple Music , through a shuffle that actually learns from how you listen, a way to split a messy playlist into something organized in one click, and an AI tool that builds you a playlist from scratch just by describing what you want.

Sift is a native macOS app I built that sits on top of your existing Apple Music library and subscription, basically a smarter client for Apple Music the same way people already use third-party apps for things like email or calendars. It’s built on Apple’s own MusicKit framework, so it talks directly to your real Apple Music library and catalog rather than faking or scraping anything. I designed and built the whole thing myself, from the interface down to the listening-history logic that actually powers the smart features, and it’s in the final stretch of App Store review right now. The core idea was pretty simple: Apple Music has a massive catalog and I already pay for it, but the actual experience of using it feels flat compared to what Spotify has trained people to expect. THis led me to I build Sift as a layer to Apple Music , through a shuffle that actually learns from how you listen, a way to split a messy playlist into something organized in one click, and an AI tool that builds you a playlist from scratch just by describing what you want.

Tools Used

Figma

Design Tool

MusicKit

Apple Music Framework

GitHub

Version Control

Xcode

IDE

Claude

LLM

Tools Used

Figma

Design Tool

MusicKit

Apple Music Framework

GitHub

Version Control

Xcode

IDE

Claude

LLM

Tools Used

Figma

Design Tool

MusicKit

Apple Music Framework

GitHub

Version Control

Xcode

IDE

Claude

LLM

Created

Created

2026

Process

Process

Spotify vs. Apple Music

Before I designed anything, I spent a while comparing Spotify and Apple Music side by side, feature by feature. Spotify wins on almost every “smart” feature: Discover Weekly, Daily Mixes, an algorithmic shuffle that actually adapts, AI DJ, all of it. Apple Music’s catalog size is right there with Spotify’s, but the software around it hasn’t caught up. I’m an Apple Music user myself because of the ecosystem and the audio quality, and when I asked around, basically every other Apple Music user I talked to said some version of the same thing: they wished Apple Music felt as smart as Spotify, but didn’t want to actually switch services to get it due to the friction of switching ecosystems.

That pain point was the main premise for Sift. Rather than build another music player, I wanted to build the layer Apple Music should have shipped to be an even more competitive service, on top of the subscription people already have.

What Sift Actually Does

The first thing I tackled was shuffle, because shuffle on Apple Music is really just a random sort, it has no memory of what you skip or replay. I built Smart Control around a single Familiarity dial that goes from deep cuts you barely play to your most-played songs, backed by individual signal weights (skips, recent plays, repeats, minutes listened) so the shuffle actually reflects real listening behavior instead of pure chance. Next to that is Auto-Sort, which takes one big messy playlist and splits it into smaller ones by genre, artist, or vibe in a couple clicks, something that used to be a manual, annoying process.

The feature I’m most proud of is the AI Playlist Generator, because it’s not something Spotify or Apple Music does either. You type a plain description, a mood, an event, an artist combo, and Sift searches all of Apple Music (not just your library) to build a real playlist out of it, and you can keep refining it with follow-up prompts until it’s right. That one feature alone is the reason I’d pick Sift over just opening Apple Music directly as it tackles a major want from music listeners. Anyone can ask an LLM to do the same but to actually have it integrated with plalist fully created and generated for usability in under a minute is something neither Spotify or Apple Music has.


Colors and fonts for a design system, made by nickelfox.com

Design Decisions

A big part of making Sift feel right was making sure it never felt like a random third-party hack bolted onto Apple Music, it needed to feel like something Apple could have shipped themselves. The clearest example of that is how reordering playlists works: when you drag a tile in the library carousel, every other tile jiggles exactly the way app icons jiggle when you rearrange your iPhone’s home screen. That one interaction does a lot of work, because the second people see it, they immediately know what to do without having to explain it.

The rest of the interface leans on the same logic. I used Apple’s glass material look throughout, translucent surfaces, soft edge highlights, blurred backgrounds, so the whole app feels native to how macOS looks right now instead of looking like a Figma file someone exported. For color, I kept it mostly dark and neutral and let one accent gradient, the same red used in the app’s icon (same red Apple Music uses as an accent) do all the work, on the Smart Shuffle toggle, the Familiarity dial, the primary buttons, so your eye always knows where the important action is without the interface causing friction. Nothing about the layout is trying to be "different" or "special", as the app is built so the actual music stays the focus.


Design Decisions

A big part of making Sift feel right was making sure it never felt like a random third-party hack bolted onto Apple Music, it needed to feel like something Apple could have shipped themselves. The clearest example of that is how reordering playlists works: when you drag a tile in the library carousel, every other tile jiggles exactly the way app icons jiggle when you rearrange your iPhone’s home screen. That one interaction does a lot of work, because the second people see it, they immediately know what to do without having to explain it.

The rest of the interface leans on the same logic. I used Apple’s glass material look throughout, translucent surfaces, soft edge highlights, blurred backgrounds, so the whole app feels native to how macOS looks right now instead of looking like a Figma file someone exported. For color, I kept it mostly dark and neutral and let one accent gradient, the same red used in the app’s icon (same red Apple Music uses as an accent) do all the work, on the Smart Shuffle toggle, the Familiarity dial, the primary buttons, so your eye always knows where the important action is without the interface causing friction. Nothing about the layout is trying to be "different" or "special", as the app is built so the actual music stays the focus.


Design Decisions

A big part of making Sift feel right was making sure it never felt like a random third-party hack bolted onto Apple Music, it needed to feel like something Apple could have shipped themselves. The clearest example of that is how reordering playlists works: when you drag a tile in the library carousel, every other tile jiggles exactly the way app icons jiggle when you rearrange your iPhone’s home screen. That one interaction does a lot of work, because the second people see it, they immediately know what to do without having to explain it.

The rest of the interface leans on the same logic. I used Apple’s glass material look throughout, translucent surfaces, soft edge highlights, blurred backgrounds, so the whole app feels native to how macOS looks right now instead of looking like a Figma file someone exported. For color, I kept it mostly dark and neutral and let one accent gradient, the same red used in the app’s icon (same red Apple Music uses as an accent) do all the work, on the Smart Shuffle toggle, the Familiarity dial, the primary buttons, so your eye always knows where the important action is without the interface causing friction. Nothing about the layout is trying to be "different" or "special", as the app is built so the actual music stays the focus.


Two iPhones displaying a Web3.0 learning app

Conclusion

While I was building the Smart Shuffle logic, I kept thinking about how Spotify handles this exact problem and came away less impressed than I expected. Spotify’s personalization works, but it’s kind of a black box since you have no idea why it’s playing what it’s playing or how to push it in a different direction. I wanted the opposite: a shuffle you could actually see and steer, where turning a dial or a signal up or down has a visible, understandable effect on what plays next. That felt like the more honest way to build a personalization feature, instead of asking people to just trust an algorithm they can’t see inside.

Beyond being a portfolio project, Sift has been a good learning experience to better develop my skills in Xcode and another project where I got to actually ship something end to end, design it, build it, debug it for real, and go through Apple’s actual submission process, not just mock it up and move on.

Conclusion

While I was building the Smart Shuffle logic, I kept thinking about how Spotify handles this exact problem and came away less impressed than I expected. Spotify’s personalization works, but it’s kind of a black box since you have no idea why it’s playing what it’s playing or how to push it in a different direction. I wanted the opposite: a shuffle you could actually see and steer, where turning a dial or a signal up or down has a visible, understandable effect on what plays next. That felt like the more honest way to build a personalization feature, instead of asking people to just trust an algorithm they can’t see inside.

Beyond being a portfolio project, Sift has been a good learning experience to better develop my skills in Xcode and another project where I got to actually ship something end to end, design it, build it, debug it for real, and go through Apple’s actual submission process, not just mock it up and move on.

Conclusion

While I was building the Smart Shuffle logic, I kept thinking about how Spotify handles this exact problem and came away less impressed than I expected. Spotify’s personalization works, but it’s kind of a black box since you have no idea why it’s playing what it’s playing or how to push it in a different direction. I wanted the opposite: a shuffle you could actually see and steer, where turning a dial or a signal up or down has a visible, understandable effect on what plays next. That felt like the more honest way to build a personalization feature, instead of asking people to just trust an algorithm they can’t see inside.

Beyond being a portfolio project, Sift has been a good learning experience to better develop my skills in Xcode and another project where I got to actually ship something end to end, design it, build it, debug it for real, and go through Apple’s actual submission process, not just mock it up and move on.

Next Project

Next Project

Next Project

Pluvus

Pluvus

Pluvus

->

->

Ronak Vusirikala © 2026

Scroll to Top

Ronak Vusirikala © 2026

Scroll to Top