Salman A.

TicketBoat required a highly scalable, automated, and secure scraping system capable of collecting millions of event listings in real-time without triggering WAF (Web Application Firewall) blocks, rate limits, or browser fingerprinting systems. The goal was to build a robust data pipeline that aggregates stadium events, concerts, and sports tickets, keeping data fresh and accurate. Our team delivered a full end-to-end web automation solution, leveraging multiple scraping technologies and orchestration systems. Objectives => Scrape millions of events across multiple ticketing websites. => Bypass advanced security layers such as CloudFront WAF, bot detection, and CAPTCHAs. => Manage proxy/IP rotation seamlessly. => Handle browser fingerprinting to mimic real user sessions. => Maintain high throughput, stability, and 24/7 uptime. => Integrate all data into TicketBoat’s internal system for real-time search. Challenges 1. Advanced WAF Protection (CloudFront) Ticketing platforms like StubHub and Ticketmaster use: => AWS CloudFront WAF => Behavioral bot detection => Device fingerprinting => Dynamic CAPTCHAs => IP reputation checks These systems aggressively block scraping traffic at scale. 2. Browser Fingerprinting Platforms detect: => Canvas fingerprinting => WebGL => AudioContext signatures => User-agent anomalies => Headless browser behavior 3. Rate Limits & IP Bans Scraping millions of pages quickly leads to: => IP throttling => Soft bans => Long-term rate limiting 4. Horizontal Scaling To scrape millions of events: => We needed parallel scraping => Distributed workers => Smart retry logic => Proxy load balancing

FastRead is an AI-driven content creation platform that turns a single thought or idea into a complete book draft using large language models. It empowers creators, authors, and entrepreneurs to generate, edit, and publish high-quality books faster combining AI creativity with a seamless user experience. My Role: As a Full-Stack & AI Engineer, I designed and implemented the platform’s Python-based AI orchestration system that coordinates multiple large language models (LLMs) including ChatGPT (OpenAI), Claude, and custom fine-tuned models for natural language generation, editing, and refinement. Key Contributions: => Engineered a multi-LLM pipeline in Python to dynamically route requests between OpenAI, Claude, and internal models for optimal creativity, tone, and factual accuracy. => Integrated speech-to-text and text-to-speech features using OpenAI Whisper and Google Speech APIs, enabling users to dictate ideas or listen to generated drafts. => Developed a FastAPI backend for book creation, content generation, and AI session management with JWT-based authentication. => Built a Next.js 15.4 frontend with TailwindCSS, Shadcn/UI, and Framer Motion, offering a modern and fluid interface for book writing and editing. => Integrated analytics tools including Hotjar, Google Analytics 4, and Facebook Pixel for behavioral tracking and conversion optimization. => Managed deployment on AWS (Lambda, EC2) with static assets stored in DigitalOcean Spaces, behind an Nginx reverse proxy for performance and security. Impact: FastRead revolutionized how creators produce long-form content reducing writing time by 70%, increasing creative consistency, and enabling real-time voice interaction with AI models. The platform now serves a growing user base of writers and educators leveraging multi-LLM creativity for book generation.