Senior Data Scientist, Buyer (Vinted, UAB)
Darbdavys: Vinted, UAB
Miestas: Kaunas
Paskelbta: 2026-10-31
💶 3433 - 6233 €/mėn. Prieš mokesčius · visa darbo diena · nuotolinis darbas: hibridinis · paskelbta 2026-10-25
Kandidatuoti šaltinyje (cv.lt) →
Job description:
The Buyer Domain at Vinted is dedicated to creating a seamless, delightful, and highly engaging experience for our members as they search, discover, and purchase items. Our challenge is unique: unlike standard e-commerce, our peer-to-peer catalog consists of hundreds of millions of entirely unique, single-inventory listings, with millions of new items uploaded daily. Connecting the right buyer with the perfect item in real-time is at the very core of Vinted's success.
To tackle this challenge at scale, we are looking to bring talented Data Scientists into the Buyer Domain's two highly specialized, ML-focused teams:
Search Relevance: This team is responsible for the retrieval and ranking systems that power search queries across all markets. Their mission is to ensure our members can effortlessly find exactly what they are looking for in Vinted's vast, unique catalog. The team's ownership and focus areas include multi-stage search ranking pipelines, query intent classification, semantic dense retrieval, and the integration of ML relevance models with content policy controls, all optimized for real-time execution.
Recommender Relevance: This team owns the Recommender System that powers Vinted’s Homepage Feed and which is responsible for driving a significant portion of all discoveries and purchases across the platform. Their mission is to inspire our members by surfacing highly personalized, engaging, and relevant recommendations. The team's ownership and focus areas include deep sequence-based user modeling, real-time candidate generation and ranking, and exploration of diverse member interests, built to handle massive scale and high-throughput constraints.
We will work with you to match your background, technical interests, and experience with the team where you can drive the most impact. Regardless of the team you join, you will be a proactive product partner. We work in a highly collaborative culture to align on technical approaches, and we validate all changes through rigorous offline evaluation and online A/B testing on our custom-built, in-house Experimentation platform.
Build and optimize end-to-end machine learning models for retrieval, ranking, and personalization (e.g. gradient boosting, two-tower architectures, sequence models, and neural rerankers).
Partner closely with platform and software engineers to serve models directly in our search and recommendation engines, optimizing for high throughput and millisecond-level latency budgets.
Lead or contribute to collaborative design proposals for new modeling approaches, features, or training pipelines.
Define tracking requirements, perform offline evaluations, and design and analyze online A/B tests to measure real-world business impact.
Collaborate with Data, Engineering and Product stakeholders to identify product opportunities, translate business problems into ML objectives, and align on technical trade-offs.
Share knowledge, participate in peer reviews, and actively…
The Buyer Domain at Vinted is dedicated to creating a seamless, delightful, and highly engaging experience for our members as they search, discover, and purchase items. Our challenge is unique: unlike standard e-commerce, our peer-to-peer catalog consists of hundreds of millions of entirely unique, single-inventory listings, with millions of new items uploaded daily. Connecting the right buyer with the perfect item in real-time is at the very core of Vinted's success.
To tackle this challenge at scale, we are looking to bring talented Data Scientists into the Buyer Domain's two highly specialized, ML-focused teams:
Search Relevance: This team is responsible for the retrieval and ranking systems that power search queries across all markets. Their mission is to ensure our members can effortlessly find exactly what they are looking for in Vinted's vast, unique catalog. The team's ownership and focus areas include multi-stage search ranking pipelines, query intent classification, semantic dense retrieval, and the integration of ML relevance models with content policy controls, all optimized for real-time execution.
Recommender Relevance: This team owns the Recommender System that powers Vinted’s Homepage Feed and which is responsible for driving a significant portion of all discoveries and purchases across the platform. Their mission is to inspire our members by surfacing highly personalized, engaging, and relevant recommendations. The team's ownership and focus areas include deep sequence-based user modeling, real-time candidate generation and ranking, and exploration of diverse member interests, built to handle massive scale and high-throughput constraints.
We will work with you to match your background, technical interests, and experience with the team where you can drive the most impact. Regardless of the team you join, you will be a proactive product partner. We work in a highly collaborative culture to align on technical approaches, and we validate all changes through rigorous offline evaluation and online A/B testing on our custom-built, in-house Experimentation platform.
Build and optimize end-to-end machine learning models for retrieval, ranking, and personalization (e.g. gradient boosting, two-tower architectures, sequence models, and neural rerankers).
Partner closely with platform and software engineers to serve models directly in our search and recommendation engines, optimizing for high throughput and millisecond-level latency budgets.
Lead or contribute to collaborative design proposals for new modeling approaches, features, or training pipelines.
Define tracking requirements, perform offline evaluations, and design and analyze online A/B tests to measure real-world business impact.
Collaborate with Data, Engineering and Product stakeholders to identify product opportunities, translate business problems into ML objectives, and align on technical trade-offs.
Share knowledge, participate in peer reviews, and actively…