Overview
TalkToTheCTO helps freelancers and contractors stop sending CVs into the void. Instead, it connects them directly with the people who actually hire: CTOs, VPs, and Heads of Engineering. Behind the scenes, an AI engine runs around the clock to find this work and package it into ready-to-use leads.
Project Context
The premise is blunt: HR rarely hires freelancers, but executives do. Therefore, the product tracks two signals at once. First, it finds companies that are actively hiring.
Second, it links each one to the real decision-maker on LinkedIn. As a result, every lead pairs proof of intent with a direct line to the person who can say yes.
Technical Architecture
I built the acquisition engine as a microservices system in Docker. Each module runs in its own container and communicates over a Redis message bus. The backend uses Python,
FastAPI, and async SQLAlchemy on top of PostgreSQL. Moreover, all AI runs locally through Ollama and a Qwen2.5 model, so there are no external API costs. A scanner then monitors
job boards, analyzes each opportunity with the model, and applies automatically.
On top of that, I built a separate marketplace on a cloud VPS. It pairs a FastAPI backend with a React and Tailwind frontend. Furthermore, it handles Stripe payments, credit-based
pricing, and a lead-exclusivity cap. Finally, a one-way sync script feeds fresh leads from the local engine to the cloud database.
Key Achievement
I deliberately split the private engine from the public marketplace into two repos. Consequently, the scraping logic stays local and protected, while the customer-facing platform
scales independently in the cloud. This clean separation keeps each side safe to change, deploy, and grow on its own.
