Context
- A growing job platform had outrun its monolith. Search and matching were the bottleneck choking growth, and the architecture had no path to the next order of magnitude of users. A rewrite was off the table; the business had to keep running.
Approach
- Sequenced the decomposition so the lights stayed on throughout.
- Carved out search and matching as independently scalable services.
- Moved analytical workloads to BigQuery for scale and cost.
What we built
- Service-based architecture replacing the monolith’s hot paths.
- A rebuilt search & matching system tuned for relevance and throughput.
- Data pipeline into BigQuery powering analytics and matching signals.
Results
- Scaled to serve 1M+ professionals.
- Search and matching got faster as the system grew, not slower.
- A platform ready for the next growth step, no rewrite required.
Stack & standards
GoDjangoVueBigQueryPostgreSQL
Secure SDLCObservability
“They cut the monolith without ever taking us down. That is the hard part, and they made it look routine.”
Verified review · Clutch 5.0
Related work