Algorithm Expert - Financial Foundation LLM

Didi-Global-; San Jose (CA); On-site; USD 180,000 - 260,000; Full time

Job summary

DiDi Global Inc. is a leading mobility technology platform seeking a senior ML researcher to drive foundation model pre-training pipelines and multi-domain sequence modeling for financial risk applications.

You will collaborate with product and engineering teams to ship high-impact models, optimize training across large-scale data, and deploy models in production with robust evaluation and monitoring.

Qualifications

• Must have: Master's degree or above in Computer Science, Mathematics, Statistics, or a related field.
• 3+ years of deep learning algorithm R&D experience with hands-on experience building a pre-trained model from scratch and completing the full training pipeline.
• Proficiency in Transformer architectures and variants (GPT/BERT/FT-Transformer/MoE).
• Familiar with PyTorch and distributed training (multi-GPU / multi-node).
• Strong experimental design skills including ablation studies and scaling-law experiments.

Responsibilities

• Foundation Model Pre-training: design and implement pre-training pipeline for financial behavior sequence models and experiments.
• Multi-source Sequence Modeling: build unified sequences across payments, ride-hailing, food delivery, and credit data.
• Ablation Studies and Evaluation Framework: create a systematic ablation framework and evaluation pipeline.
• Production Deployment: integrate pre-trained representations into downstream risk-control scenarios and deploy online.

Skills

Deep learning

Transformer architectures

PyTorch

Distributed training

Sequence modeling

Education

Master's degree or above in Computer Science, Mathematics, Statistics, or a related field

Tools

PyTorch

Multi-GPU / multi-node

Job description

Company Overview

DiDi Global Inc. is the world’s leading mobility technology platform. It offers a wide range of app-based services across markets including Asia-Pacific, Latin America and Africa, including ride hailing, taxi hailing, chauffeur, hitch and other forms of shared mobility as well as auto solutions, food delivery, intra-city freight, and financial services.

DiDi provides car owners, drivers, and delivery partners with flexible work and income opportunities. It is committed to collaborating with policymakers, the taxi industry, the automobile industry and the communities to solve the world’s transportation, environmental and employment challenges through the use of AI technology and localized smart transportation innovations. DiDi strives to create better life experiences and greater social value, by building a safe, inclusive and sustainable transportation and local services ecosystem for cities of the future.

For more information, please visit: www.didiglobal.com/news

Team Overview

DiDi Fintech is deeply involved in the Latin American market, actively deploying a variety of financial services such as electronic payment, loan, credit cards, and merchant acquiring. We aim to provide local users with more convenient and high-quality services.Cost-effective financial solutions. Payment Risk Algorithm team is a core pillar of DiDi’s global Fintech business, building ML‑driven risk systems for fast‑growing Latin‑American payment and wallet scenarios. We design end‑to‑end fraud detection models to balance transaction security and user payment experience, defending against payment fraud, account compromise and emerging adversarial risks across overseas markets. You will collaborate closely with product, engineering and regional business teams to ship high‑impact algorithm capabilities that power financial expansion.

Role Responsibilities

• Foundation Model Pre-training: Design and implement the pre-training pipeline for financial behavior sequence foundation models, including pre-training objective selection (CLM/MLM/Hybrid), Tokenization architecture experimentation (Flat/3D-Transformer/KVT), and scaling experiments.
• Multi-source Sequence Modeling: Build a unified sequence representation for behavioral data across multiple domains (payments, ride-hailing, food delivery, credit); design and validate the impact of data-source mixing ratios on model performance. Multi-entity and Multi-scale Fusion: Design an Account-Card dual-dimension sequence modeling scheme, along with a cross-temporal-scale fusion architecture bridging micro-level behaviors (millisecond-granularity event tracking) and macro-level behaviors (day/week-level transactions).
• Ablation Studies and Evaluation Framework: Build a systematic ablation experiment framework; design a freeze-backbone + linear head evaluation pipeline to drive architecture decisions.
• Production Deployment: Integrate pre-trained representations into downstream risk-control scenarios (stolen-card detection, credit scoring, etc.); design a Blending Module and complete SFT fine-tuning and online deployment.

Role Qualifications

Must Have:

• Master's degree or above in Computer Science, Mathematics, Statistics, or a related field.
• 3+ years of deep learning algorithm R&D experience, with hands-on experience building a pre-trained model from scratch and completing the full training pipeline.
• Proficiency in Transformer architectures and variants (GPT/BERT/FT-Transformer/MoE), with practical sequence modeling experience.
• Familiar with at least one mainstream deep learning framework (PyTorch preferred); experience with distributed training (multi-GPU / multi-node).
• Solid experimental design skills: ability to independently conduct ablation studies and scaling-law experiments and draw reliable conclusions.
• Strong engineering implementation skills; able to iterate efficiently on model code and training pipelines.

Nice to Have:

• Modeling experience in financial risk control / anti-fraud / credit scoring.
• Publications on Foundation Models / Self-Supervised Learning (NeurIPS/ICML/ICLR/KDD/WWW, etc.).
• Familiarity with Contrastive Learning, ELECTRA, cross-modal fusion, and related techniques.
• Experience with time-series / event-sequence modeling (e.g., TimeMixer, TrajGPT).
• Experience with graph neural networks or graph-based anomaly detection.

EEO Statement

• We create customer value – We strive to always create valuable experiences for our users in everything we do. Our focus is to always innovate new experiences that are safe, pleasant, and efficient.
• We are data-driven – We are strong believers in making informed decisions, that’s why we are data-driven. We can better navigate the business landscape strategically by analyzing valuable metrics.
• We believe in Win-win Collaboration – Success is a team sport. When we work to help our partners and colleagues win, we win, too. While keeping everyone's best interest at heart, we communicate with candor and execute with excellence in all we do.
• We believe in integrity – Integrity is at the very core of our business. We are people who always want to do the right thing. Our intentions are sincere, we speak our minds and listen to each other.
• We always strive to do better. That means venturing beyond our comfort zones, learning from our mistakes, and helping each other grow.
• We believe in Diversity and Inclusion – Diversity is one of our biggest strengths. Our differences are what makes us distinct. We respect each other and believe in equal opportunities for all.

We are committed to building inclusive and diverse teams.

At DiDi, we believe that our differences are our biggest source of strength. That‘s why we are committed to promoting equal opportunities to all candidates and employees as an Equal Opportunity Employer.

Employment and advancement decisions at DiDi are always made based on the needs of the position and the qualifications of the candidate. We do not discriminate against any employee or applicant based on their gender, age, sexual orientation, nationality, marital status, pregnancy/maternity, disability, race, religion and beliefs, or any other status protected by applicable laws wherever we operate.

We are committed to building inclusive and diverse teams, and a workplace that is free from discrimination and harassment, because that’s how we create better products and services, make better decisions and better serve the communities we’re a part of.

I acknowledge that prior to submitting this application, I have read and accepted the Privacy Notice for Candidates which is available on https://careers.didiglobal.com/terms