AKP / PROJECTS

Projects

7 things built for a hackathon, coursework, and the products he wanted to exist. Most link to source.

  • BuilderBridge screenshot

    BuilderBridge

    AI-assisted construction control room that connects schedules, critical-path planning, weekly commitments, field progress, RFIs, and project documents. Its project-aware agent cites live data, prepares reviewable changes, and applies them only after permission checks and human approval.

    Next.js, React, TypeScript, Tailwind CSS, Vercel AI SDK, Prisma, PostgreSQL, Better Auth, Supabase

  • SynergySphere screenshot

    SynergySphere

    Team workspace built at the Odoo x Paradox hackathon at IIT Madras: task management, threaded chat, mentions, notifications, and project dashboards.

    Next.js, TypeScript, Tailwind CSS, shadcn/ui, better-auth, PostgreSQL, Drizzle

  • Claso screenshot

    Claso

    Full-stack code assistant that uses custom-trained transformer models to write code comments and context-aware commit messages for Python and JavaScript, without calling an external model API.

    React, TypeScript, Tailwind CSS, Monaco Editor, Python, FastAPI, PyTorch, Transformers

  • Ramso screenshot

    Ramso

    One-click AI video generator that turns a concept into an educational explainer: a script, Manim animation, a voiceover, and a downloadable MP4.

    Python, Streamlit, Agno, OpenAI, Manim, Neets AI, FFmpeg

  • TypeScript

    IntelCraft

    Outreach assistant for students and freshers. It looks up a company, drafts a personalized email or cover letter, and streams progress over WebSockets.

    React, TypeScript, FastAPI, PostgreSQL, WebSockets, LangGraph, Vercel, Railway

  • Python

    Todo Telegram Agent

    Telegram reminder bot that turns a sentence into a scheduled task. Groq parses the intent and time, SQLite stores the task, and a scheduler sends the reminder with Done and Snooze actions.

    Python, FastAPI, SQLModel, SQLite, python-telegram-bot, APScheduler, Groq

  • TypeScript

    YT-Slice

    Calculates the total and average length of a YouTube playlist range, and how long that range takes at different playback speeds.

    Next.js, TypeScript, Tailwind CSS, shadcn/ui, React Hook Form, YouTube Data API

Simplicity is prerequisite for reliability.

– Edsger W. Dijkstra