Overview
Trakr is our most ambitious in-house project to date: an intelligence layer for the job hunt. Built from the ground up, it gives candidates one place to manage a career move and automates the tedious work of tracking applications and their status.
The challenge
Processing thousands of unstructured emails and application notifications across disparate platforms while maintaining near-perfect accuracy in status detection. The system had to be fast, secure, and capable of handling complex authentication flows with major email providers.
The solution
We developed a multi-model classification engine on Python and FastAPI, which runs live Gmail data through our own ML models to automatically detect and categorize job statuses. The entire ecosystem is delivered via a high-performance native iOS application built with Flutter, utilizing Redis for real-time processing and GCP for secure, scalable infrastructure.
Outcome
Trakr is in closed beta, where early testers report a 90% reduction in time spent tracking applications by hand.
