Shiraz Ali
Full Stack Dev | Agentic AI & Automation Systems
Profile in progress
Shiraz is building their profile!
Ustaad.ai (http://Ustaad.ai) is a production-deployed, multi-agent Pakistani AI tutoring platform designed to make quality education accessible to every student in Pakistan — in the language they actually think in.
Each subject has a dedicated AI tutor persona (Master Bilal for Science, Master Asad for Math, Apa Sara for English, Master Sahab for Urdu) that teaches conversationally in Roman Urdu and English — the way Pakistani students naturally communicate.
Key Features:
Subject-specific multi-agent architecture — each tutor is a separate AI agent
Real-time AI chat sessions with voice input support
Built-in quiz system (Lo Quiz) for instant practice
Learning streak tracker to build study habits
Syllabus settings per subject
Clean, modern dashboard UI
Tech Stack: Next.js · FastAPI · Google Gemini API · Docker · Google Cloud Run · Supabase
Built for: Google Cloud AI Seekho 2026 Competition
0
2
Built a production-ready, end-to-end agentic AI system designed to automate the entire B2B lead generation and cold outreach pipeline — from discovery to inbox.
The system uses the Google Places API with dynamic neighborhood-level scanning to find local businesses that have no website. Each lead is then deeply enriched using Firecrawl, which scrapes Yelp, Facebook, Instagram, and LinkedIn to find business owners, contact emails, and detailed business context.
From there, a custom multi-agent pipeline powered by Google Gemini Flash (with OpenRouter failover) analyzes each business, identifies specific pain points a website would solve, scores them on a 1–10 opportunity scale, and generates a hyper-personalized cold outreach email tailored to their exact situation.
Everything is tracked in real-time through a modern Next.js dashboard using Server-Sent Events (SSE), with one-click options to review, edit, regenerate in different tones (Friendly, Professional, Casual), and send emails directly via Resend.
2
28