AI Document Understanding Co‑Op — Remote/Hybrid Data Science

Ancestry; United States; Hybrid; USD 34,000 - 48,000; Part time

Job summary

Ancestry is seeking a Data Science Co-Op to join the Content AI team, focusing on developing AI models that extract and organize information from historical records. You will build, train, and fine-tune models to surface insights that connect people to their ancestors.

You will work with engineering teams to train, optimize, and deploy models for product development and content creation across our family history platform. This role requires a strong background in NLP and multi-modal modeling.

Qualifications

• Pursuing an advanced degree (Master's or PhD preferred) in a quantitative field.
• Specialization in generative AI & LLMs, embeddings, LoRA/QLoRA, vector databases, transformers, NLP.
• Strong Python skills and experience with NLP tools and libraries.

Responsibilities

• Innovate with state-of-the-art AI for Document Understanding tasks (OCR, handwriting recognition, transcription, NER, RE, summarization, knowledge graphs).
• Analyze and optimize multi-modal models in zero-/few-shot scenarios.
• Collaborate on cloud deployment with ML Ops and Data Science Engineers.
• Communicate findings and proposed solutions to technical and non-technical audiences.

Skills

Python

NLP

Transformers

LangChain

Multi‑modal models

Education

Master's or PhD preferred in CS/DS/Statistics/Math/Linguistics

Tools

Hugging Face Transformers

LangGraph

Google Gemini API

Vertex AI

AWS SageMaker

Job description

Ancestry is seeking a Data Science Co-Op to join the Content AI team, focusing on developing AI models that extract and organize information from historical records. You will build, train, and fine-tune models to surface insights that connect people to their ancestors.

You will work with engineering teams to train, optimize, and deploy models for product development and content creation across our family history platform. This role requires a strong background in NLP and multi-modal modeling.