Skip to content
View UditBuilds's full-sized avatar
🎯
Focusing
🎯
Focusing

Block or report UditBuilds

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
UditBuilds/README.md

Udit Kumar

AI Agent & Automation Engineer — Solo Founder, Delhi

I build production-grade AI agents and automation systems, and I run a D2C brand that some of them power. Not tutorials. Not demos. Real tools solving real problems — most of them mine first.

🎯 Open to remote AI engineering roles — agent development, automation, intelligent systems.


🟢 Live in Production

Project What it does Stack
Prism AI-native spaced-repetition PWA — SM-2 scheduling, LLaMA flashcard generation, PDF→cards, push notifications Next.js 14 · TypeScript · Supabase · Groq Live ↗
The Glam Shelf Live D2C false-eyelash brand on Shopify — 142+ orders, powered by the GlamShelf Twin agent Shopify · Razorpay · WATI Live ↗

🛠 Built & Open Source

Real projects with full source available — built to solve specific problems end-to-end.

Project What it does Stack
GlamShelf Twin AI customer-support agent powering my own D2C brand — handles ~80% of WhatsApp/Instagram DMs autonomously, escalates the rest via Telegram Flask · Claude API · WATI · Shopify · Telegram
AI Intelligence Daily 90-second daily AI briefing bot delivered to Telegram every morning — curated for founders Python · Groq LLaMA 3.3-70B · GitHub Actions
Daily Digest Newsbot Twice-daily AI-summarized news briefing to Telegram — free, RSS-backed, multi-source aggregation Python · Groq LLaMA 3.3-70B · GitHub Actions
VidSage YouTube → structured study notes — Whisper transcription, chunked audio for 3hr+ videos, PDF/Markdown export Flask · Groq Whisper · yt-dlp
DraftTwin Multi-tenant AI DM-drafting SaaS for D2C founders — Flask-Login auth + per-brand "brain" system Flask · Groq LLaMA 3.3-70B · SQLite · Railway

🧠 How I Build

  • Agents that ship — classification routing (AUTO / DRAFT / ESCALATE), human-in-the-loop approval, brand-voice grounding, real webhooks and payment/fulfilment integrations.
  • Cost-aware LLM engineering — multi-key rotation, provider fallback (Groq / Claude / Gemini), exponential backoff, and response deduplication.
  • End-to-end ownership — from Supabase schema and PWA service workers to Railway / Render / Vercel deploys and the CI that keeps them running.

🧰 Stack

Languages — Python · TypeScript · SQL AI — Claude API · Groq (LLaMA 3.3-70B, Whisper) · Gemini · SM-2 Web — Next.js 14 · Flask · React · Tailwind CSS Data — Supabase · PostgreSQL · SQLite Infra — Railway · Render · Vercel · GitHub Actions · n8n Commerce — Shopify · WATI · Razorpay


📍 Currently

  • Running The Glam Shelf — a D2C false-eyelash brand on Shopify (142+ orders and growing)
  • Building Prism in the open — an AI-native learning tool, shipped across 12+ sessions
  • Open to remote AI roles — agent development, automation, intelligent systems

📫 LinkedIn · Instagram

Pinned Loading

  1. glamshelf-twin glamshelf-twin Public

    AI-powered customer support twin for a live D2C beauty brand — built with Flask, Claude API, WATI, Instagram webhooks, Shopify, and Telegram. Handles 80% of customer DMs autonomously.

    Python 3

  2. Prism-Productivity-Tool Prism-Productivity-Tool Public

    AI-native spaced-repetition PWA — SM-2 scheduling, LLaMA flashcard generation, PDF→cards, focus timer, and push notifications. Built with Next.js 14, TypeScript & Supabase.

    TypeScript

  3. AI-Intelligence-Daily AI-Intelligence-Daily Public

    Daily AI intelligence brief for founders. Python + Groq LLaMA 3.3-70B + GitHub Actions + cron-job.org. Delivers 90-second AI briefing to Telegram every morning.

    Python

  4. agentgrade agentgrade Public

    Open-source evaluation framework for AI customer support agents. 4-dimension LLM-as-judge (tone, accuracy, compliance, resolution), provider-agnostic, configurable rubric. Dogfooded on GlamShelf Twin.

    Python