Case study · Ongoing
Lootstack
A multi-game collectibles platform that brings card discovery, recognition, pricing, collections and local trading into one product.
Context
One product across the collecting workflow.
Card catalogues vary across games, printings, languages, finishes and markets. Prices keep moving, while collectors expect search, recognition and portfolio value to agree. Lootstack turns those parts into one coherent experience across web and mobile.
Search and filter cards across supported games, sets and printings.
Scan a physical card and resolve the relevant catalogue entry.
Manage collections and decks with market-aware values.
Find relevant matches and coordinate trades around local stores.
Data foundation
Keeping the catalogue usable as the inputs change.
Data arrives from multiple external sources with different identifiers, structures and update cycles. Ingestion and review workflows normalize records, reconcile variants, enrich the catalogue, apply market data and publish changes to the customer product.
Applied ML
Recognition designed around the moment of use.
The scanner connects camera input to the catalogue and the collection workflow. The scope spans model training and evaluation, mobile inference, handling ambiguous printings and the product experience around confirmation and capture.
- 01Find the card
Detect the physical card in a live camera view.
- 02Resolve the printing
Match against the relevant catalogue and distinguish close candidates.
- 03Keep the user in control
Confirm uncertainty instead of silently committing the wrong result.
- 04Continue the workflow
Add the result to a collection or make it available for trade.
Scope
Scope across the whole product.
Customer-facing features and the systems underneath are treated as one product. That keeps decisions connected from the catalogue and recognition flow through to delivery, measurement and operations.
Customer journeys, interaction design, web and mobile applications, collections, decks and trading workflows.
Ingestion, ETL, catalogue modelling, pricing, search, review workflows and data quality.
Recognition research, training and evaluation pipelines, mobile inference and scanner integration.
Backend services, realtime features, administration, billing, communications, reliability and delivery.
Have a product with moving parts?
Make it feel
like one product.
Tell us what customers need, what the underlying systems must coordinate and where the risk sits. We’ll give you a direct read on the path forward.
Start the conversation