About the role
Responsibilities:
- Own and develop match-making, recommendation, ranking, and personalisation systems.
- Work on creating a novel, real-time adaptive matchmaking engine that learns from user interactions and other signals.
- Design ranking and recommendation algorithms that make each user's feed feel curated for them.
- Build user embedding systems, similarity models, and graph-based match scoring frameworks.
- Explore and integrate cold-start solutions.
- Partner with Data + Product + Backend teams to deliver great customer experiences.
- Deploy models to production using fast iteration loops, model registries, and observability tooling.
Requirements:
- 3-5 years of experience working on personalisation, recommendations, search, or ranking at scale.
- Prior experience in a B2C product - social, ecommerce, fashion, dating, gaming, or video platforms.
- Exposure to a wide range of popular recommendation and personalisation techniques, including collaborative filtering, deep retrieval models (e. g., two-tower), learning-to-rank, embeddings with ANN search, and LLM approaches for sparse data personalisation.
- Can train models AND ship them - experience with end-to-end ML pipelines.
- Understand offline and online evaluation, A/B testing, and metric alignment.
- Experience with vector search, graph-based algorithms, and LLM-based approaches is a big plus.