Info Kerja Ketapang - Youth Employment Platform
Role & duration: Digital Platform Consultant (part-time), technical owner and project lead. September 2026 to present, for Save the Children Indonesia, as part of a youth employment programme in Ketapang Regency, West Kalimantan. In progress.
TL;DR
- Problem: A job-matching platform for local young people existed, but it had serious security gaps, data problems, and no clear owner. The programme needed it to be reliable before launch.
- Solution: Audited the existing app, fixed the worst problems first, kept the codebase instead of rebuilding, and organised delivery across developers, the client team, and the local manpower office.
- Status: 47 development issues, most of them closed. A beta is live and has been presented to the local manpower office, with a joint evaluation planned. Launch is targeted for early 2027.
- Tech Stack: Laravel, React, MySQL, Tailwind CSS.
Context & Problem
The platform connects job seekers in the regency (vocational school students and early school leavers) with local employers. It is one part of a larger youth empowerment programme, so its job is to produce placements the programme can report on, not just to exist.
When I joined, the app had been built but not owned. Nobody was accountable for its stability, its security, or what it would take to launch it.
- Who had the problem. The programme team, which needs a platform it can put in front of the local government and rely on. And the job seekers and employers, who would be handing over personal data.
- Why it mattered. A job platform for young people holds CVs, contact details, and employment history. A leak at launch would hurt the people the programme exists to help and would end the partnership’s credibility with the local government.
- Condition before. A first technical audit found 117 findings across a 192-item checklist: 45 critical, 58 medium, 14 low. Among the critical ones were problems that should never reach production, including debug output left on in a live environment and a source repository that was publicly reachable.
Constraint
- Part-time, with delegated development. I could not write every fix myself. Most of the building is done by other developers, so my leverage is deciding what gets built, in what order, and to what standard.
- Fixed contract end. The consulting contract ends before the platform is fully launched. Everything I decide has to leave the team able to run it without me.
- Several stakeholders, different needs. The programme team wants placement numbers, and the local manpower office wants workforce planning data on top of daily vacancies.
- New requests mid-project. The local government’s requests arrived after the original specification. Each one was a trade-off between schedule, budget, and scope.
Decision & Trade-off
Fix the existing app, do not rebuild. Rebuilding would have been cleaner and slower. The audit showed that most of the problems were fixable in place, and the programme could not spend the time or money on a rewrite. I took the existing codebase, fixed the critical findings first, and put the remaining work in order of risk.
Security before features. It is tempting to show progress through new screens. I held feature work behind the critical audit findings, and moved the repository to private as soon as the audit showed it was publicly reachable. A launch with fewer features is recoverable. A launch with a leak is not.
Make the rules visible in the product. The partnership agreement requires employers to give a reason when they reject an applicant. I made that a required step in the app, so the rule is enforced by the product and not by reminders.
Rename before launch. The original name was close to an existing government site. It was flagged in a weekly meeting and the platform was renamed before any public launch, which is much cheaper than renaming after people know it.
Plan the handover from the start. I act as interim admin until the end of 2026, then the platform passes to the programme’s youth centre. A single admin is a single point of failure, so the plan is to train more than one person from the youth centre while I still run it.
Implementation
Audit and triage. Worked from a 192-item checklist and ranked it by what could hurt users first. About 100 items were closed by mid-October, with the critical ones prioritised.
Core platform. Login and registration (including Google sign-in), job seeker and employer profiles, vacancy posting and verification, applications, applicant selection, interview scheduling, dashboards for admins and employers, a public landing page, and automatic CV generation in an ATS-friendly format.
Safety and accountability. Activity logs, two-factor authentication for admins, and an archive process.
Requests from the local manpower office. Employer registration data, a workforce data section, a public company page, an admin recap with Excel export, and a next-year workforce planning module that lets the office see what employers expect to hire.
Delivery process. The beta is updated only from the main branch, and weekly feedback from the programme team is turned into specific issues so nothing sits as a vague request.
What is still open
The platform is not launched, so I do not claim outcomes it has not produced yet. The open work is notifications, scheduled emails, a final pre-launch security pass, and the decision on the launch date against the end of my contract. Those are the items I am tracking toward early 2027.
What I Learned
- Security comes before visible progress. New screens make a project look further along than it is. The audit findings decided the order of work, not the demo schedule.
- Scope requests are trade-offs, not favours. The manpower office’s planning module came after the original specification. It is valuable, and it also competes with launch work for the same time and budget, so it needs an explicit decision.
- Leading is different from building. In earlier roles I was the sole engineer. Here the work is deciding, delegating, and reviewing, and I measure progress by what the team ships.