Onsite_Mid/Sen Data Scientist (Không yêu cầu tiếng anh)

Onsite Bank

Scope of work

Summary:

The Data Scientist role focuses on building propensity models and sizing market opportunities for VMO project with banking clients. This hybrid position requires a blend of data science and engineering skills to create production-grade pipelines. The role is pivotal in driving sales strategy through data-driven outputs. Candidates should have a strong background in Python and SQL, along with proven experience in propensity modeling.

Propensity Modeling & Ranking:

• Predict CTR and engagement probability
• Use models like XGBoost / LightGBM
• Optimize ranking using NDCG, MRR, Precision@K

Feature Engineering:

• Build features from user behavior, interaction signals, and context
• Handle sparse and noisy data

Cold Start Handling:

• Design strategies for new users and new offers
• Implement hybrid and fallback approaches

Bias Handling:

• Mitigate popularity and exposure bias
• Implement diversity and re-ranking strategies

Model Evaluation:

• Define and track CTR, conversion, NDCG, AUC
• Continuously improve engagement metrics

Scalability & Deployment:

• Build systems for millions of users
• Enable real-time or near real-time inference

Requirements

Key Skills:

• Strong Python - the existing model suite is Python-heavy
• Advanced SQL large complex datasets
• Proven propensity modelling experience - built and delivered, not just theoretical
• Genuine hybrid capability - data science and data engineering
• Confident communicator with non-technical stakeholders Hands-on with XGBoost, LightGBM

Nice to Have

• Experience with MLOps and model deployment
• Experience with AWS SageMaker
• A/B testing and experimentation knowledge