Required stack
필요 기술
ReactVue.jsSvelteNode.jsPythonFlaskFastAPIMachine Learning (ML)SQLDockerGitHTMLCSSJavaScriptSoftware ArchitectureWeb DevelopmentBootstrapSpring Boot
Project brief
프로젝트 내용
I want to build a web-based expense manager that focuses on one crucial ability: automatically placing every transaction into the right category using AI/ML. No mobile clients are needed right now; the entire experience lives in the browser.
Here’s the workflow I have in mind. I type in, paste, or import a bank CSV; the system predicts the category—groceries, rent, travel, etc.—with high accuracy. Whenever I correct a mistake the model should quietly learn, getting smarter the more I use it. A clean dashboard will then let me filter, search, and see totals by category. I do not need budget tracking or receipt scanning in this first release.
The tech stack is open. React, Vue, or Svelte on the front end are all fine, and I’m equally happy with a Python (Flask/FastAPI), Node, or similar back end as long as the ML component integrates smoothly. Feel free to reuse public datasets or lightweight models; what matters is that the pipeline is reproducible and the code is well organised.
Deliverables
• Responsive web UI with secure sign-up / login
• AI/ML service that auto-categorises expenses and updates itself from user corrections
• Dashboard showing per-category totals, search, export to CSV
• Clear documentation and setup script (Docker or similar)
Acceptance criteria
• ≥80 % categorisation accuracy on the supplied test file
• Page loads under 2 s on a standard connection
• Codebase lint-clean with basic unit tests passing
Ship it in a Git repo with instructions and I’ll handle deployment from there.
Here’s the workflow I have in mind. I type in, paste, or import a bank CSV; the system predicts the category—groceries, rent, travel, etc.—with high accuracy. Whenever I correct a mistake the model should quietly learn, getting smarter the more I use it. A clean dashboard will then let me filter, search, and see totals by category. I do not need budget tracking or receipt scanning in this first release.
The tech stack is open. React, Vue, or Svelte on the front end are all fine, and I’m equally happy with a Python (Flask/FastAPI), Node, or similar back end as long as the ML component integrates smoothly. Feel free to reuse public datasets or lightweight models; what matters is that the pipeline is reproducible and the code is well organised.
Deliverables
• Responsive web UI with secure sign-up / login
• AI/ML service that auto-categorises expenses and updates itself from user corrections
• Dashboard showing per-category totals, search, export to CSV
• Clear documentation and setup script (Docker or similar)
Acceptance criteria
• ≥80 % categorisation accuracy on the supplied test file
• Page loads under 2 s on a standard connection
• Codebase lint-clean with basic unit tests passing
Ship it in a Git repo with instructions and I’ll handle deployment from there.