KerjaHub

The Problem

The job market in Southeast Asia is fragmented. Traditional job portals optimize for volume — quantity of listings and applications — which buries qualified candidates and drowns employers in mismatched applicants.

KerjaHub's thesis: match on skills and readiness, not keywords, and help candidates become job-ready instead of just job-seeking.

Where It Stands

KerjaHub is deliberately still at the architecture-and-design stage ahead of go-to-market. The current work is the unglamorous part that determines whether the platform can scale later:

• Defining service boundaries so the platform can grow without a rewrite
• Designing the data model for matching, assessments, and employer workflows
• Selecting the stack for day-one productivity and long-term performance

Architecture Decisions

As CTO, I own the technical direction. The decisions so far optimize for a small team shipping fast without painting the platform into a corner:

Current stack direction:

• Golang services: high-performance core APIs with clear service boundaries
• Next.js web: SEO-critical job listings and the employer dashboard
• Bun tooling: fast runtime for supporting services and tooling
• PostgreSQL + Redis: relational core with caching where matching gets hot

Why Architecture First

Scaling Ezzi School taught me what it costs to retrofit foundations under a growing product. With KerjaHub the order is reversed: service boundaries, data model, and delivery pipeline first — then go-to-market with a platform that won't need a rewrite at its first growth spike.