DEV Community: Aryan

🚀 Day 4 | AI Gym Assistant 💪 - Project Progress Update!

Aryan — Mon, 16 Jun 2025 09:35:34 +0000

Today was nothing short of a breakthrough! 🔥

I integrated a drag-and-drop upload experience powered with buttery-smooth GSAP animations, making it not just functional but fun to interact with!
From the moment a user drops a gym equipment image — our backend kicks in with:
✅ Google Vision API for equipment detection
✅ Gemini AI for suggesting beginner-friendly workouts
✅ Real-time GIFs via GIPHY that bring each exercise to life! 🎥💥

💡 imagine this: You upload a simple photo of a dumbbell... and instantly get animated workout suggestions tailored to that equipment — GIFs, muscle groups, tips and all!

What I worked on today:
🧠 Refined the prompt engineering for better AI output
🎨 Improved the front-end styling and responsiveness
🎞️ GIFs now feel like something out of a top-tier fitness app

Still to come:
➡️ User authentication
➡️ History tracking for uploaded images
➡️ Sharing your AI-generated routine with friends!

🔥 Day 3 of building something really cool!

Aryan — Sat, 14 Jun 2025 16:32:52 +0000

I’ve been working on a smart fitness project that I’m super excited about — I didn’t post about it earlier (was deep in the build mode the last 2 days 😅) — but imagine this: upload a photo of any gym equipment and instantly get 2–3 beginner-friendly exercises, complete with videos, tips, and targeted muscle info! 💪📸

No fitness knowledge? No problem.
This tool is built for beginners who walk into a gym and think:
"What can I even do with this machine?"
**
🚧 What I’ve built so far (in just 3 days):**
✅ Detect gym equipment from a photo using Google Cloud Vision API
✅ Use Gemini (Google’s LLM) to generate exercises, GIF/video links, muscle info, and beginner tips
✅ Created a backend with Node.js + Express to connect the dots
✅ Parsed and cleaned messy AI output into clean JSON (trust me, this was tricky 😅)
✅ Ready to hook this into a simple, clean React frontend

Next up: Designing the UI to actually show this magic ✨

If you're into fitness, AI, LLMs, or just love seeing creative tools come to life — I’d love to hear your thoughts! 💬
Let’s connect and build cool stuff together 🚀

Day 5: Dynamic Routing & Header Support

Aryan — Tue, 10 Jun 2025 12:39:10 +0000

🤔 Ever wondered how servers handle dynamic routes like /user/:id or validate headers?

Today, I built dynamic routing and header support from scratch in Node.js—no Express, just raw HTTP.

How it works:
🚀 Trie-Based Routing

Efficiently maps paths like /posts/:id using a Trie structure.

No more brittle if-else or regex chaos!

🛡️ Header Parsing & Validation

Extracts Content-Type, Authorization, and other headers.

Enforces rules (e.g., Accept: application/json).

📦 Modular Design

Clean separation: Parser → Router → Response Handler.

Ready to scale beyond hardcoded routes.

DAY 4 – Streaming Body Parse

Aryan — Mon, 09 Jun 2025 10:01:46 +0000

🤷‍♂️Ever wondered how large file uploads don’t crash servers?

Today, I built a streaming body parser from scratch using raw Node.js TCP sockets — no Express, no frameworks.

--> Accumulate data chunks from the TCP socket
--> Parse HTTP request headers to find Content-Length
--> Read request body safely, waiting for full body based on length
--> Support plain text and JSON parsing for POST requests
--> Avoid memory leaks by handling data in streams rather than buffering everything at once.

Most developers rely on express.json() or middleware like multer for uploads. I wanted to understand what really happens under the hood.

This deep dive helped me build a custom HTTP server that correctly handles POST request bodies — the backbone for handling forms, APIs, and uploads — without third-party dependencies.

Day 3 – The Thought That Changed Everything

Aryan — Sun, 08 Jun 2025 10:31:48 +0000

🧠 Ever wondered what really happens before http.createServer() in Node?

I did — and decided to dig way deeper.

While working on my server recently, a thought struck me:

Can I build this from even lower down? Like, handle the raw socket connections myself — understand the handshake, data transfer at the byte level, and have full control over every bit sent and received.

So that’s exactly what I started doing:

✅ Handling raw TCP data streams
✅ Parsing HTTP methods, paths, and headers manually
✅ Building a response engine without relying on the http module
✅ Rewriting my Trie-based router to work directly on top of TCP sockets

Most devs just use express() or http.createServer().
But I wanted to know — what’s really going on underneath?

This deep dive is teaching me:
→ How TCP connections work from the ground up
→ How net.Socket behaves internally
→ How to build custom servers, proxies, or mini frameworks from scratch

Here is the Github Gist for the sample of the new implementation :

DAY 2 – Trie-Based Router Like Fastify

Aryan — Sat, 07 Jun 2025 09:34:33 +0000

🌲Fastify (alternative of express) uses a Trie for route matching — today, I built my own.

✅ Each URL segment is a node in the tree

✅ Lookup is O(k) for k segments

✅ Supports /user/:id and wildcards

✅ Cleaner and faster than if-else chains

I never thought a data structure would make routing this efficient.

Frameworks do this behind the scenes — I wanted to understand the “why”.

Tomorrow I’ll move to a body parser that works in chunks.

This is a Github Gist u can see the implementation here. It's a small portion of the code

Day 1 — “Why I’m Building My Own HTTP Server from Scratch”

Aryan — Fri, 06 Jun 2025 18:53:27 +0000

Most backend devs use Express or Fastify.
I decided to build my own HTTP server — from scratch, using raw Node.js.

No frameworks. No shortcuts.
Just sockets, buffers, and my brain.

Why? Because I want to understand how servers really work under the hood.

In the next few days, I’ll share how I’m building this:

A blazing-fast router using Trie
Streaming request body parser
JWT auth
Custom middleware engine
Reverse proxy

Let’s see how far I can take this.