Financial Foundation LLM Architect

DiDi Global; San Jose (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

DiDi Global in San Jose seeks a Senior ML Research Engineer to lead foundation-model pre-training for financial behavior sequences, exploring objective choices and tokenization architectures. You will collaborate with product, engineering and regional teams to ship high-impact risk and payments models.

The role emphasizes multi-source sequence modeling across payments, ride-hailing, food delivery and credit; you will design cross-domain fusion, ablation studies, and scalable training pipelines,

Qualifications

• Master's degree in CS/Math/Statistics or related field.

Responsibilities

• Design and implement pre-training pipelines for financial behavior models.
• Build a unified sequence representation across payments, ride-hailing, food delivery and credit.
• Develop ablation studies and evaluation frameworks to guide architecture decisions.
• Integrate pre-trained representations into risk-control scenarios and online deployment.

Skills

Transformer architectures

PyTorch

Distributed training

Experimental design

Model evaluation

Education

Master's degree or above in Computer Science/Math/Stats

Tools

PyTorch

Job description

DiDi Global in San Jose seeks a Senior ML Research Engineer to lead foundation-model pre-training for financial behavior sequences, exploring objective choices and tokenization architectures. You will collaborate with product, engineering and regional teams to ship high-impact risk and payments models.

The role emphasizes multi-source sequence modeling across payments, ride-hailing, food delivery and credit; you will design cross-domain fusion, ablation studies, and scalable training pipelines,