From May 24 th, 2026, Fairleap AI will transition to new domains as 🌐 Website:fairleap.faizath.com (formerly fairleap.cloud) |
An AI co-pilot for the financial well-being of Indonesia's gig drivers.
Fairleap AI is an AI-powered financial well-being platform built for Gojek / GoTo ride-hailing and delivery drivers. It turns a driver's raw, irregular trip data into clear earnings insights, personalized financial guidance, and wellness support — helping drivers move from day-to-day income uncertainty toward long-term financial stability.
The name says it all: a fair leap forward for the gig workers who keep Indonesia moving.
🥇 Fairleap AI was built and submitted to the Alibaba Cloud Indonesia GenAI Hackathon 2025, where it was selected as a Top 12 Finalist.
The Alibaba Cloud Indonesia GenAI Hackathon 2025 is a national-scale Generative AI hackathon held on May 24–25, 2025 in Jakarta, Indonesia (Habitate, Kuningan). It is organized by Alibaba Cloud, the GoTo Group, and KOMDIGI (Indonesia's Ministry of Communication and Digital Affairs).
The competition challenges innovators across Indonesia to build impactful, real-world solutions using Generative AI on Alibaba Cloud — leveraging services such as the Qwen large language models via DashScope. It brings together developers, students, and startups to prototype GenAI applications that address tangible problems, with finalists pitching their solutions to a panel of industry judges.
Out of all participating teams, Fairleap AI advanced to the Top 12 Finalists, recognizing its use of Generative AI to tackle income uncertainty and well-being for Indonesia's gig-economy drivers.
 |  |  |
 |  |  |
 |  |  |
 |  |  |
 |  |  |
 |  |  |
 |  |  |
 |  |  |
 |  |  |
 |  |  |
Gig-economy drivers are the backbone of Indonesia's on-demand economy, yet they face structural challenges that traditional financial tools ignore:
- Income uncertainty — earnings swing daily with demand, weather, surge pricing, and hours worked, making it nearly impossible to budget or plan ahead.
- No financial visibility — drivers rarely have tools to understand why they earn what they earn, or how to optimize their working hours.
- Limited access to financial advice — conventional financial advisory and investment products are not designed for variable, cash-based gig income.
- Well-being is overlooked — long hours on the road create fatigue and stress that directly affect both safety and earning capacity.
Fairleap AI combines machine-learning forecasting with Generative AI (RAG-powered) advisory to give every driver a personal financial and wellness assistant:
- 📊 Earnings Analytics & Forecasting — ML regression models (XGBoost / scikit-learn) analyze historical trip data to surface earning patterns and predict future earnings, so drivers know what to expect and when to drive.
- 🤖 AI Financial Advisor — a Generative AI advisor (powered by Alibaba Cloud Qwen via DashScope, with RAG grounding) delivers personalized financial tips tailored to each driver's income profile.
- 📈 Investment Guidance — an AI investment bot recommends suitable, easy-to-understand investment options for variable gig income.
- 🧘 Wellness Companion — an AI wellness assistant provides health, fatigue, and work-life-balance guidance to keep drivers safe and sustainable.
- 💬 Conversational Chatbot — a RAG-based chatbot answers driver questions in natural language, grounded in Fairleap's domain knowledge.
- 🖥️ Unified Dashboard — a clean web dashboard brings earnings, analytics, financial advice, and wellness together in one place.
| Component | Description | Tech Stack | Repository | Deployment URL / Hugging Face Repo |
|---|---|---|---|---|
| 🌐 Frontend | Next.js web app & driver dashboard | Next.js 15, React 19, TypeScript, Tailwind CSS, shadcn/ui, Recharts, GSAP | Fairleap-AI/fairleap-fe | fairleap.faizath.com(formerly fairleap.cloud) |
| ⚙️ API Backend | Core REST API, auth & data layer | Node.js, Express 4, MongoDB (Mongoose), JWT, Passport, Nodemailer, Docker | Fairleap-AI/fairleap-api | fairleap-api.faizath.com(formerly api.fairleap.cloud) |
| 🧠 AI / RAG Backend | ML forecasting & GenAI advisory services | Python, Flask, Gunicorn, XGBoost, scikit-learn, pandas, NumPy, DashScope (Qwen) | Fairleap-AI/fairleap-ai | Served with Docker internally |
| 📚 SFT Dataset | 42,743 synthetic Indonesian driver conversations | JSONL, Parquet, PyArrow, Hugging Face Datasets | Fairleap-AI/fairleap-driver-chat-sft-43k | 🤗 datasets/fairleap-ai/fairleap-driver-chat-sft-43k |
| 🤖 Qwen3.5-4B Adapter | QLoRA adapter for driver chat & tool calling | PEFT (QLoRA), Unsloth, TRL, Transformers, bitsandbytes, PyTorch | Fairleap-AI/fairleap-v1-clm-qwen3.5-4b-adapter | 🤗 fairleap-ai/fairleap-v1-clm-qwen3.5-4b-adapter |
| 🤖 Sahabat-AI 8B Adapter | QLoRA adapter on the Llama-3 Indonesian base | PEFT (QLoRA), Unsloth, TRL, Transformers, bitsandbytes, PyTorch | Fairleap-AI/fairleap-v1-clm-sahabatai-8b-adapter | 🤗 fairleap-ai/fairleap-v1-clm-sahabatai-8b-adapter |
This project is licensed under the MIT License.