Finding and reaching the right businesses shouldn't take a team.
Sales and outreach teams typically stitch together several tools — a scraper here, a CRM there, a separate outreach platform — to find leads and get in front of them. LeadBoba needed to do all of it in one place: discover businesses in any city or industry, surface verified contact details, and generate outreach that doesn't read like a template. And it needed to reach real customers organically, without a marketing team behind it.
What's actually running under the hood.
Large-scale data extraction
A crawling system that discovers businesses worldwide and collects company name, address, phone, email, website, reviews, ratings, and social profiles — routed through rotating proxy infrastructure to keep large-scale crawling stable and reliable.
Async processing at scale
Crawling and enrichment run as background jobs on dedicated worker servers, coordinated through Upstash Redis, so the app stays responsive while heavy scraping and data processing happen off the main request path.
AI-powered outreach
LLM-driven cold email generation with custom prompt engineering and retry handling, plus an AI writing-assist layer that refines grammar, tone, and clarity before a message goes out.
Production-grade infrastructure
API server on AWS ECS, workers on Hetzner, images in Amazon ECR, monitoring via CloudWatch, health checks through an Application Load Balancer, and internal SSL — the same infrastructure discipline that goes into every client build.
Secure by default
JWT and HTTP-only cookies, bcrypt password hashing, Google OAuth, and environment configuration stored securely in AWS S3 for dynamic, per-service task configuration.
150,000 pages, one developer, no ad budget.
Instead of relying on paid acquisition, LeadBoba's growth is driven by a programmatic SEO engine I built to generate targeted landing pages automatically — one for every meaningful city-and-industry combination, each with its own structured data and search-intent-matched copy. That system is now live with over 150,000 indexed pages.
It's this combination — real infrastructure experience plus a working understanding of what actually gets a page found — that carries over into every client site I build. A website that looks right but was never built to rank is a missed opportunity; I build for both from day one.
Want something similar?