Skiable
Flight search and exploration platform
THE CHALLENGE
A flight search tool has to handle complex queries, return structured results quickly, and let people compare many options.
TECH STACK
Engineering Approach
Fast Search
Built the search with Next.js and TypeScript, with data structures chosen for fast filtering.
Server-Side API Routes
Wrote API routes in Next.js that query MongoDB and return structured results.
Complex Interaction Design
Designed responsive interfaces for multi-step forms, date ranges, passenger counts, and result comparison.
Modular Architecture
Organized the code with reusable hooks and clear component boundaries.
Key Technical Decisions
The choices that shaped the architecture and how easy it is to maintain.
Next.js API routes
Used API routes instead of a separate backend.
One deployment, and TypeScript types shared between the frontend and the backend.
MongoDB aggregation
Used the aggregation pipeline for search.
Complex filtering and sorting is more efficient in the database than in application code.
Custom hooks for search state
Wrapped search and filter state in reusable hooks.
Complex logic lives in composable units that several components can use.
Challenges & Solutions
Searches with several filters were slow.
Added compound indexes in MongoDB for the common filter combinations.
The search form's state was hard to manage and debug.
Centralized it in a search context with TypeScript types and validation at each step.
Comparing many flight options was hard.
Built a sortable table with persistent filters.
Outcomes
Skiable is a flight search platform with filtering, sorting, and a responsive interface across devices.
KEY TAKEAWAYS
Search interfaces are more complex than they look. A simple form hides many edge cases and performance concerns.
