Description
Grab is Southeast Asia's leading superapp, providing services such as food delivery, financial management, and ride-hailing.
The Fulfilment Tech Family builds the systems that power Grab's marketplaces across Southeast Asia, designing real-time, distributed systems and Machine Learning (ML) solutions that process hundreds of millions of requests each day.
We are looking for a Senior Principal Data Scientist to lead our shift toward automated marketplace optimization, advancing how we use data and ML to automate pricing, dispatch, and supply management decisions.
Responsibilities
- Design and implement end-to-end RL systems that combine model-based RL, offline RL, simulation, and online learning into a unified training pipeline.
- Build latent world models and marketplace state representations that capture supply-demand interactions, location-based patterns, and behavioural signals from users and drivers.
- Develop systems that optimize across multiple marketplace levers simultaneously,pricing, dispatching, and supply repositioning,to expand the set of achievable outcomes for the business.
- Create policy evaluation frameworks and establish monitoring systems that allow safe deployment of new decision-making policies in production.
- Fine-tune open-source large language models on domain-specific data to build capabilities for prediction, reasoning, and simulation within marketplace applications.
- Design and implement training strategies for language models, including supervised fine-tuning, preference-based alignment, and iterative improvement methods.
- Work with data engineers and backend engineers to integrate RL and LLM systems into real-time production environments serving millions of users.
Qualifications
- PhD in Computer Science, Operations Research, Applied Mathematics, or related field with at least 10 years of experience.
- Proficiency in RL fundamentals, including Markov Decision Processes, stochastic control, and reward design trade-offs.
- Experience building production ML/ RL systems with online learning or simulation-based optimization.
- Knowledge in world models and sequential modelling, including latent dynamic models (RNNs, transformers, state-space models) and representation learning for complex systems.
- Hands-on experience fine-tuning large language models in production, including supervised fine-tuning and at least one preference-based alignment method (RLHF, DPO, or GRPO).
- Experience with distributed training frameworks, such as DeepSpeed, FSDP, or Megatron-LM.
- Proficiency in Python and ML frameworks (PyTorch or TensorFlow).
- Experience with scalable computing platforms, such as Spark or Ray.
Benefits
- Term Life Insurance and comprehensive Medical Insurance.
- GrabFlex, a benefits package that suits your needs and aspirations.
- Parental and Birthday leave, and Love-all-Serve-all (LASA) volunteering leave.
- Confidential Grabber Assistance Programme.
- FlexWork arrangements, such as differentiated hours.