We are looking for a Principal / Head-level builder who has built recommendation systems from scratch and scaled them in production. This role is for someone who loves being deep in data, modeling, and production code, and has owned recommender stacks that served millions of users.
You will be responsible for building our core recommendation engine - candidate generation, ranking, and personalization.
What You Will Build & Own
1. Recommendation Engine Development
- Build end-to-end recommendation systems from ground up: data pipelines, candidate generation, ranking, re-ranking and serving
- Implement and improve collaborative filtering, content-based, two-tower, sequential and graph-based models
- Build robust solutions for cold-start, diversity, and exploration vs exploitation
2. Production Implementation
- Take models from prototype to production with a focus on latency, scalability and reliability
- Build real-time and batch inference pipelines, feature stores, and monitoring for model drift and performance
- Work hands-on with data and ML engineers to deploy and optimize large-scale systems
3. Ranking & Personalization
- Develop advanced ranking models using learning-to-rank, multi-task learning and deep learning
- Improve relevance, engagement and conversion through continuous experimentation
- Define and own offline evaluation [NDCG, Recall@K, MRR] and online evaluation [A/B tests, interleaving]
4. Team & Technical Leadership
- Be the go-to expert for all things recommendations - set coding, modeling and evaluation standards
- Mentor data scientists and ML engineers and lead technical design reviews
- Lead by building - this is a 70% hands-on IC + 30% technical leadership role
What You Bring
- 8+ years in Data Science / Machine Learning with 4+ years specifically building recommendation systems in production
- You have shipped at least 2 large-scale recsys that served live users and can show measurable lift in CTR, conversion, retention or engagement
- Deep expertise in retrieval and ranking, embeddings, and sequential modeling
- Strong programming in Python, strong SQL, and experience with PyTorch / TensorFlow and Spark
- Experience with vector databases, ANN search [FAISS][HNSW], and model serving at scale[ScaNN]
- Strong background in A/B testing and analysis of large-scale behavioral data
- Track record of leading small teams of 3-8 scientists / engineers while remaining hands-on
Bonus if you have:
- Built recsys for marketplace, e-commerce, content or OTT
- Experience with real-time personalization, session-based recommendations
- Experience with Vespa, Elasticsearch, or similar ranking infrastructure
