AI MLOps Platform Frontend
Frontend for the management console of a Kubernetes MLOps platform (Wishket fixed-term engagement as a senior developer, 122 days)
As a senior developer, I built the frontend of the management console for an MLOps platform that manages AI training jobs and GPU resources on Kubernetes. I joined through a fixed-term engagement on Wishket (a Korean IT outsourcing marketplace) for 122 days, working remotely four days a week.
What I did
- Built the management console frontend for a Kubernetes MLOps platform with Next.js (App Router) and TypeScript
- Built dashboards and charts for monitoring GPU node and cluster resource usage
- Built resource management screens for single and distributed training jobs, workspaces, image registries and more
- Implemented a Monaco-based YAML editor for Kubernetes resources and real-time notifications over SSE
- Built a design system of shared UI components with Storybook
Scale
- 159 screens, and a design system of 119 shared components with 27 Storybook stories
- An infrastructure management console covering clusters, node groups, MIG, PV/PVC and network policies
Tech stack
- Next.js 14 App Router, TypeScript, TanStack Query and Table, Jotai
- Radix UI, NextUI, Recharts, Monaco Editor, Storybook, Jest