Staff Machine Learning Engineer (Research Scientist) - DFAI

Responsibilities

• Owning the end-to-end technical strategy for a foundation model built on one of the world's richest financial datasets, from pretraining architecture to production serving.
• Doing research that ships: driving decisions from experimentation through production systems that serve real customers and power multiple product teams.
• Working across the full ML stack, including pretraining objectives, architecture design, distributed training, serving infrastructure, monitoring, and cross-team integration.
• Setting technical direction and mentoring a high-caliber team, with your work amplifying the capabilities of engineers and product teams across Plaid.
• Helping hundreds of millions of consumers achieve greater financial freedom through the ML capabilities you build and ship.

Qualifications

• MS: 7–12+ years of industry experience with a demonstrated track record of technical leadership and production delivery.
• PhD: 5–9+ years of industry experience with evidence of technical leadership (tech lead, principal/staff-equivalent roles) and end-to-end production ownership.
• Prior technical leadership experience (tech lead, principal, or staff) with demonstrated cross-team influence and mentorship.
• Deep expertise in Transformers/LLMs/Foundation Models, including large-scale training or domain adaptation.
• End-to-end production ownership; proven track record shipping models through training, serving, monitoring, and iteration in live environments.
• Distributed training experience and strong Python + software engineering fundamentals at a staff level.
• Able to drive technical alignment across teams: setting standards, defining integration patterns, and influencing beyond your immediate scope.
• Fintech / financial data domain experience - Nice to have.
• External publications or open-source contributions - Nice to have.
• Experience defining ML platform capabilities (serving infra, feature stores) used across multiple teams - Nice to have.