Job Title Data Scientist – Recommendation Systems
Location Bangalore
Experience 3–8 years (flexible based on depth in ML systems)
Job Description
We are looking for a Data Scientist (Recommendations) to design, build, and scale personalized recommendation systems that power discovery, ranking, and user engagement across our products.
Requirements
Key Responsibilities
Recommendation & ML Design and develop recommendation systems including:
- Collaborative Filtering (user-item, item-item) Content-based and hybrid recommenders
- Ranking and re-ranking models Embedding-based retrieval (ANN, vector search)
- Train, evaluate, and iterate on models using offline metrics (NDCG, MAP, Recall@K) and online A/B experiments Production ML & Systems Optimize inference for scale (caching, batching, approximate nearest neighbors)
- Build real-time and batch recommendation pipelines
- Monitor model performance, data drift, and system health
Data & Experimentation
- Work with large-scale datasets (clicks, impressions, transactions)
- Define success metrics for recommendations (CTR, CVR, retention)
Collaboration
- Work closely with product, data, and backend teams to translate business problems into ML solutions
- Contribute to ML best practices, documentation, and system design
Required Skills
Core ML
- Strong understanding of: Recommendation algorithms Ranking and learning-to-rank
- Embeddings and similarity search
- Experience with Python and ML libraries (PyTorch / TensorFlow / Scikit-learn)
- Data & Systems Strong SQL skills; experience with large datasets
- Familiarity with vector databases / ANN libraries (FAISS, ScaNN, Elasticsearch/OpenSearch KNN, Milvus)
Good to Have
- Experience with: Search or feed ranking systems
- Real-time recommendations
- Knowledge of: MLOps tools (MLflow, Airflow)
- Experience in e-commerce, ads, content platforms or marketplaces
What You'll Work On
- Personalized home feeds and search ranking "People also viewed" recommendations
- Cold-start and long-tail problems
- Large-scale experimentation and model optimization
Nice Behavioral Traits
- Strong problem-solving and system-thinking mindset
- Ability to balance model quality vs production constraints