Senior Data Scientist, Fraud & Trust ML

Grab; Selangor; On-site; MYR 180,000 - 280,000; Full time

Job summary

Grab Malaysia seeks an experienced machine learning professional to combat fraud by analyzing transactional data and deploying ML models at scale. Report to the Data Science Manager II as an IC in a full-time on-site role at our Petaling Jaya office, growing expertise in agentic systems.

You will push ML and LLM boundaries, collaborating with product and software teams to ensure seamless integration of detection systems while advancing risk strategies across the Grab ecosystem.

Qualifications

• Degree in a quantitative field (CS/Math/Physics/Engineering).
• 3+ years of Python/SQL with large-scale datasets.
• Experience building and deploying ML models with TF, PyTorch, XGBoost, or similar.

Responsibilities

• Partner with teams to architect scalable DS solutions integrating ML and agentic LLMs.
• Research latest algorithms and generative AI to counter fraud.
• Prepare and augment diverse data to build high-fidelity training datasets including synthetic data via LLMs.
• Design, train and fine-tune model architectures using GNNs, Transformers, and boosted trees.
• Lead end-to-end lifecycle from deployment to monitoring with engineering teams.
• Leverage Generative AI tools to accelerate coding, testing, and productivity.

Skills

Python

SQL

Spark

TensorFlow

PyTorch

XGBoost

LightGBM

Scikit-learn

Transformers

RAG

LangChain

LangGraph

Claude

ML Deployment

Education

Bachelor's degree in a quantitative field

Tools

LangGraph

LangChain

Claude

Job description

Grab Malaysia seeks an experienced machine learning professional to combat fraud by analyzing transactional data and deploying ML models at scale. Report to the Data Science Manager II as an IC in a full-time on-site role at our Petaling Jaya office, growing expertise in agentic systems.

You will push ML and LLM boundaries, collaborating with product and software teams to ensure seamless integration of detection systems while advancing risk strategies across the Grab ecosystem.