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

Data

Python

Data Science

Amazon

Testing

Banking

Scalability

Data Engineering

Experimentation

XGBoost

LightGBM

MLOps

AWS SageMaker

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