Machine Learning Engineer (research Scientist) - Dfai

Plaid Plaid · Fintech · San Francisco, CA · All Departments

Research Scientist role focused on advancing Plaid's foundation models by developing novel architectures, pretraining objectives, and fine-tuning strategies. The role involves working across the full ML stack from data and feature engineering to training pipelines, model serving, and monitoring, with a strong emphasis on shipping research into production systems for financial applications.

What you'd actually do

  1. Building a foundation model on one of the world’s richest financial datasets that no one else has.
  2. Doing research that ships: moving from experimentation and prototypes to production systems serving real customers.
  3. Working across the full ML stack, from pretraining objectives and architectures to serving infrastructure and monitoring.
  4. Collaborating with a high-caliber team and seeing your work amplify the capabilities of multiple product teams.
  5. Helping hundreds of millions of consumers achieve greater financial freedom through data-driven products.

Skills

Required

  • MS or PhD in ML/AI/CS/Stats/Applied Math
  • 1-3 years of industry experience building and deploying ML models
  • Distributed training experience
  • strong Python + software engineering fundamentals

Nice to have

  • PhD preferred
  • Fintech / financial data domain experience
  • External publications or open-source contributions

What the JD emphasized

  • evidence of both research depth and production delivery
  • Strong applied ML research skills with production delivery experience
  • Depth in Transformers/LLMs, representation learning, or large-scale model training
  • Demonstrated ability to ship models to production (not just prototype)

Other signals

  • building foundation models
  • pretraining datasets
  • train and evaluate foundation models
  • production-grade serving and monitoring systems
  • novel model architectures
  • pretraining objectives
  • fine-tuning strategies
  • production machine learning systems
  • model serving
  • monitoring
  • evaluation frameworks
  • adapt foundation models to solve specific business challenges
  • translate research advances into production capabilities