Production planning for sea-based aquaculture
Erlvin is a production planning system that helps planners build, validate, and optimize production cycles — from smolt supply and stocking through growth, harvest goals, and fallowing. It combines a visual planning grid with a calculation engine that projects biomass, reconciles plans with Fishtalk actuals, tracks MTB limits, and checks your plan against regulatory and operational rules. Erlvin grew from the convergence of several streams: the Plan Land-based project's unified data model and architecture, knowledge and patterns from Fishtalk — both the codebase and the documentation project — the Fishtalk APIs and their work on Batch and other domain models, and UI and product insights from Håkon and Jørgen, who helped shape what the product is today.
Every component, every algorithm, every line of code in Erlvin was written by AI agents — not a single line was hand-typed by a developer. This is a fundamentally new way of building software: instead of manually translating requirements into code, we describe what we need and the AI builds it. The cycle is fast, the iterations are cheap, and the bottleneck shifts from coding capacity to domain understanding.
That's where AKVA Group sourcing comes in. We're crowd-sourcing domain expertise from everyone who touches aquaculture planning — planners, support engineers, product people, biologists, managers. When you share your experience and opinions on the Feedback page, it flows directly into the AI development loop. This isn't a suggestion box that gets triaged quarterly — your input reshapes what gets built next and how the AI reasons about the domain.
Start by filling out a short profile — just a sentence or two about your background and how you relate to production planning. Do not include personal data — no full names, no email addresses, no GDPR-sensitive information. The AI doesn't need to know who you are. It needs to understand your vantage point so it can weigh and contextualize your feedback.
I've worked at AKVA group for 15 years. I'm not a planner myself, but I support customers on a regular basis with Fishtalk's planning module and see the same pain points come up again and again.