BikinKarya - AI Work Experience Simulator
An AI-powered platform that helps fresh graduates and final-year students build real work experience through simulated project briefs, kanban task management, and a recruiter-ready portfolio.
Problem & Objective
Indonesia faces a growing employment paradox: thousands of educated graduates remain unemployed while job vacancies go unfilled due to skill mismatch. Fresh graduates have no real work experience to show, portfolios filled with academic assignments are ignored by HRD, and the gap between university curricula and industry expectations continues to widen.
BikinKarya was built to close that gap by simulating an actual work experience from start to finish. Users receive AI-generated project briefs that feel like real client work, manage their tasks on a kanban board, and build a public portfolio from their completed outputs, giving them something concrete to show recruiters without waiting for an internship opportunity.
Challenges & Solutions
1. Generating Briefs That Feel Like Real Client Work
Generic AI prompts produced briefs that felt academic and unrealistic. I engineered domain-specific prompt templates for UI/UX Design, Graphic Design, and Digital Marketing, each structured with client background, business goals, constraints, and deliverables, calibrated per difficulty level so Junior briefs are execution-focused while Senior briefs introduce strategic ambiguity and stakeholder complexity.
2. Deploying to Google Cloud Run with Docker
The competition required deployment to Google Cloud Run, which meant containerizing a Next.js app with Supabase Storage and external API dependencies. I configured a multi-stage Dockerfile to keep the image lean, handled environment variable injection at runtime, and resolved cold start latency by tuning the minimum instance count.
3. Portfolio Upload Flow After Task Completion
The transition from marking a task done to uploading portfolio items needed to feel seamless without interrupting the user's focus. I designed a modal-based upload flow triggered automatically when a card moves to Done, with Supabase Storage handling thumbnail uploads and URL validation ensuring result links are properly formatted before saving.
Results & Impact
- Delivered a complete MVP with Study Case Generator, Kanban Board, Portfolio Builder, and Interview Simulator within the competition timeline.
- Each completed brief produces a public portfolio item shareable directly to recruiters via a unique URL.
- The difficulty-tiered brief system covers Junior through Senior complexity, making the platform useful across different readiness levels.
- Successfully containerized and deployed to Google Cloud Run as required by the competition, with Supabase handling both database and file storage.
Specific Role
Full Stack Developer & Product Designer
Owned the entire product from PRD to deployment. Designed the system architecture, built the Next.js frontend and API routes, engineered the Gemini prompt templates, set up Prisma with Supabase PostgreSQL, configured Docker and Google Cloud Run deployment, and designed the UI from scratch.
Tech Stack
Status
Completed / Active