In-game asset marketplace with learning-to-rank infrastructure
Eldorado.gg operates a peer-to-peer marketplace for in-game currencies, accounts, and items, processing trades across major gaming communities. The tech stack reveals a data-driven ranking operation: Python, SQL, BigQuery, dbt, and deployed machine-learning models (MLflow, Feast, Tecton) powering a learning-to-rank system. Active projects around ranking signal A/B testing and scalability, combined with documented pain around weak heuristic ranking, indicate they're moving from rule-based matching toward learned ranking — a shift typical of marketplaces scaling past manual curation.
Eldorado.gg is a marketplace platform enabling gamers to buy and sell in-game digital assets — gold, accounts, items, and services — with built-in trade protection. Founded in 2018 in Vilnius, Lithuania, the company operates a two-sided network matching sellers of earned in-game assets with buyers seeking to accelerate progression or monetize playtime. The product surface includes a checkout flow under active iteration and a backend ranking stack designed to surface relevant listings. The team is currently 51–200 people, hiring across marketing, HR, data, engineering, product, and security roles in the Vilnius office.
Python, SQL, BigQuery, dbt, Grafana, Amplitude for analytics; MLflow, Feast, Tecton for machine learning; Angular for frontend. They use TikTok, Twitch, and YouTube for performance tracking.
Yes. Active roles span engineering (1), data (1), product (1), and security (1), plus marketing (5), HR (2), and management (2). Hiring is concentrated in Vilnius, Lithuania.
A scalable ranking stack using learning-to-rank models, A/B testing of ranking signals, checkout flow improvements, and employer branding campaigns. Key challenges include weak heuristic ranking, lack of data instrumentation, and fraud prevention.
Other companies in the same industry, closest in size