Lead Data Scientist – Document AI & Multimodal ML

Next Frontier Capital; Tampa (FL); On-site; USD 140,000 - 200,000; Full time

Benefits offered by this job

Competitive compensation

Health benefits

On-site wellness centers

Retirement savings plan

Job summary

JPMorganChase is seeking a Data Scientist Lead to oversee a team focused on image classification, text categorization, and intelligent data extraction from scanned documents. You will guide ML lifecycle work from prototyping to production deployment on AWS EKS, leveraging Python, PyTorch, TensorFlow, and Hugging Face transformers.

Preferred candidates hold 7+ years in data science with 2+ years in document AI or NLP, and possess strong stats, CV, NLP, and OCR expertise.

Qualifications

• Bachelor’s or higher in a quantitative field (CS, math, OR, data science).
• 7+ years in data science/quant analytics with 2+ years in document AI, CV or NLP.
• Strong stats, modeling, and Python proficiency with ML/DL tooling.

Responsibilities

• Lead and mentor data scientists in advanced analytics for document AI.
• Define analytical strategy for document understanding using CV, NLP, and multimodal methods.
• Develop and deploy multimodal document understanding and text categorization models.
• Design experiments, data quality, annotation strategy, and active learning.
• Prepare data, select models, and deploy on AWS SageMaker and EKS production pipelines.
• Collaborate with data/ML engineers to integrate models into document processing workflows.

Skills

Python

PyTorch

TensorFlow

AWS SageMaker

AWS Bedrock

NLP

Computer Vision

OCR

SQL

Oracle

Java

Groovy

Transformers

Multimodal

Education

Bachelor's degree

MS/PhD

Tools

OpenCV

scikit-learn

pandas

NumPy

matplotlib

seaborn

Job description

JPMorganChase is seeking a Data Scientist Lead to oversee a team focused on image classification, text categorization, and intelligent data extraction from scanned documents. You will guide ML lifecycle work from prototyping to production deployment on AWS EKS, leveraging Python, PyTorch, TensorFlow, and Hugging Face transformers.

Preferred candidates hold 7+ years in data science with 2+ years in document AI or NLP, and possess strong stats, CV, NLP, and OCR expertise.