Sr DS(Personalisation) (NCS/Job/ 3728)

Build end-to-end recommender systems:
o
Candidate generation (retrieval models, embeddings, ANN search)
o
Ranking / re-ranking (GBDT, deep learning, hybrid approaches)
•
Design objective functions aligned with business goals (CVR, revenue, margin)
•
Develop decision systems (multi-objective optimization, constraints)
•
Implement and analyze online experiments (A/B, multi-armed bandits)
•
Define feature pipelines (user embeddings, sequence features, context signals)
Current challenges
•
Transition from batch recommendations → near real-time personalization
•
Multi-objective optimization (e.g., relevance vs margin vs inventory)
•
Cold start and sparse user behavior
Tech stack (indicative)
•
Python, SQL
•
TensorFlow / PyTorch / XGBoost / LightGBM
•
Feature stores, batch + streaming pipelines
What we’re looking for
•
Strong experience in recommender systems or ranking problems
•
Deep understanding of ML systems (not just modeling)
•
Experience with experimentation and causal thinking
•