Things I built for myself

When I need a tool that doesn’t exist or costs way too much, I just build it. Every app here started as a real annoyance in my own workflow, solved by vibe-coding with AI.

Built with
  • Claude
  • GPT Codex
  • Cursor

More projects

Web app TypeScript

SEOAlter

Keyword research from live search data, not stale databases.

Mobile web app TypeScript

My Wallet

A phone-first tracker for project income, dues and spending.

From my old notebooks & pitches

Early ideas & thought experiments

Long before I had a team or AI coding tools, I used to sketch inventions, write concept docs, and pitch wild ideas to my professors just to share how we could solve everyday problems. Here are a few straight from my unorganized archive.

  • High school notebook sketch

    Virtual Workstation

    Imagining cloud computing before I even knew the term existed.

    The story

    Back in 2017, I had never heard the phrase "cloud computing." I just kept wondering why phones and budget devices had to lag and drain their batteries doing heavy work. So I drew up a diagram called Virtual Work Station: a phone sends heavy data wirelessly to a high-spec remote computer, that machine crunches everything, and sends the finished result right back. Only later did I realize the tech world already had a name for my sketch.

    1. 1Phone sends heavy task over Wi-Fi
    2. 2High-spec remote workstation processes it
    3. 3Finished output returns without draining battery
  • Hardware thought experiment

    Tele Pad (Haptic Insole)

    An Arduino shoe insole that communicates silently through vibration zones.

    The story

    A playful hardware concept from my student days: I wondered how you could send simple signals without a screen or sound. I designed a diagram for a shoe insole wired to an Arduino wireless circuit and four tiny vibration motors at the left and right toe and heel, mapped to A, B, C, and D, triggered from a smartphone. I never built it, of course, but designing the circuit got me hooked on how software can talk to the physical world.

    1. 1Smartphone app sends a 4-option wireless signal
    2. 2Arduino receiver inside the shoe decodes it
    3. 3Toe and heel motors vibrate for A, B, C, or D
  • 1st-year university pitch to my professor

    Smart Shelf (স্মার্ট শেলফ)

    A search-to-light physical rack so pharmacies and stores never lose an item.

    The story

    In my first year at Green University, I pitched an idea called Smart Shelf to my professor. Pharmacies and shops waste so much time digging through hundreds of tiny drawers. My design paired a physical shelf grid (labeled A-0001, B-0002, and so on) with LED indicators and a laptop search app. Type a medicine name on the computer, and the exact drawer lights up right in front of you.

    1. 1Search item name on the laptop app
    2. 2Database matches the shelf coordinate (e.g. A-0001)
    3. 3LED indicator lights up on the exact compartment
  • System concept & proposal

    Project R: 3D Depth Road Map

    Using phone gyroscopes to crowdsource potholes and broken roads in 3D.

    The story

    Google Maps is great for 2D routes and live traffic, but it never warns you if the road ahead is full of potholes or unmarked speed breakers. I planned a system that works just like traffic detection: when multiple cars hit the same bad stretch, their phone gyroscopes register the exact same shock spike. By combining those readings across drivers, the map renders road quality in 3D, coloring smooth roads green and broken stretches red.

    1. 1Phone gyroscopes detect road shocks while driving
    2. 2Repeated bumps at the same GPS spot confirm a pothole
    3. 3Live 3D map marks smooth roads green and broken ones red
  • Figma prototype & AI concept doc

    AI Outfit Picker & Daily Helpers

    Three everyday AI apps I designed in Figma when LLMs first took off.

    The story

    Right after shipping my first AI plugin in 2023, I designed an app concept in Figma called Gentop. It combined three daily helpers I wished I had: an AI Outfit Picker that checks the local weather, your day’s schedule, and your style to pick what to wear; a plain-language song lyrics explainer; and a food ingredient chatbot that checks product labels against your personal health conditions.

    1. 1Checks weather, daily plans, or food labels
    2. 2Matches against your personal profile and health needs
    3. 3Gives instant, plain-language daily recommendations

Client work at CodeMyPixel

Alongside my own tools, I plan and lead products for clients. A few of them:

  • TicketUAI ticket aggregator, GlobalProject manager
  • VidalSignsHealthTech AI, United StatesProject manager & system design
  • QuasarSEOWordPress plugin, NetherlandsLead developer
  • SocialQuasarSocial media automation, NetherlandsTechnical project manager & architect
  • TopFloor TrendsContent AI, United KingdomTechnical project manager
  • TextGPT + IQR.CodesKnowledge management, GlobalProject manager & AI specialist
See client projects in detail

Need a tool like this for your business?

CodeMyPixel builds custom apps, AI agents and automation. Tell me what slows your team down.

Start a project