FishVerify is a fish identification and regulation app for anglers — live scan or photo import, local rules, marine weather, and catch logging. I worked on the Android client and the Java API that powers it, and set up Google AutoML model training so species detection stayed accurate while cutting response time in the field.
- Android
- Java API
- Google AutoML
- Image recognition
- GPS & catch log
- Regulations data
The problem
Anglers need instant species ID plus trustworthy bag limits, size slots, and seasons for the water they are standing in. On-device capture and a remote recognition pipeline both have to feel snappy — slow model responses kill trust when someone is holding a fish over the water.
What I delivered
- Android application work Client flows for scan/import identification, species details, regulations, and related product surfaces tied to the live API.
- Java API services Backend endpoints that serve identification results, species profiles, and regulation data to the mobile apps.
- Google AutoML training setup Model training pipeline for fish detection that improved latency so the app could return matches faster without sacrificing recognition quality.
- Field-ready product loop Identification paired with regulations, weather context, and catch logging so the recognition result is immediately useful.
Outcome
fishverify.com presents the live product across App Store, Google Play, and Galaxy Store. Faster AutoML-backed detection made the Android experience feel responsive enough for real trips, not just demos.