All projects

02 / WEB DEVELOPMENT

TripSift

Travel planning with filters, comparisons, and ranked suggestions that explain why a destination fits.

Visit live site
● ● ●tripsift.online↗
Public interface of TripSift
PUBLIC INTERFACE
The public Norwegian homepage, captured in October 2026.

The objective

Make it easier to choose a destination based on budget, time, and interests. A questionnaire leads to ranked suggestions, while country guides, itineraries, and comparisons provide context for the decision.

My contribution

  • Developed the product idea and defined travel planning, filtering, and comparison features.
  • Worked on design decisions, usability, testing, and troubleshooting over several iterations.

How I worked

I used Codex and other AI tools extensively for code implementation. My work focused on developing the idea, defining functionality, making design choices, testing, troubleshooting, and improving the product over several iterations. The technical overview describes the implementation; it does not imply that I wrote all the code independently.

Technical challenges & approach

Travel suggestions with an explanation

A rule-based scoring model combines budget, interests, season, duration, transport, and climate. Results include reasons and downsides alongside the score. Repository Vitest tests examine determinism and how budget changes affect ranking, among other cases.

Keeping choices consistent across pages

TypeScript models, separate feature modules, and next-intl connect the questionnaire, results, and multilingual pages. Local storage supports saved choices. The code also includes optional Supabase authentication, local-data migration, and SQL policies for user-owned data; these account features have not been verified in the public demo.

Learning focus

The project provides a practical way to understand how data models, filtering, state, and translations affect usability. A ranking is more useful when people can understand both its reasoning and its limits.

Project scope

Travel inspiration with illustrative, manually maintained cost estimates. No live prices or booking. The public quick-search-to-results journey was checked. Test code was reviewed, but project tests and account features were not executed as part of this portfolio review.

NEXT PROJECT

TrainPilot