Wanted: Post‑doc to Teach Machines to Read Historical Names

 Overview

La Rochelle Université is seeking a Post‑doctoral Researcher in Computer Science, specializing in Historical Handwritten Text Recognition (HTR). The position offers the opportunity to develop and refine systems for handwriting recognition in historical documents, leveraging machine learning, deep learning, and textual analysis.

Roles & Responsibilities

The post-doc will tackle tasks such as:

• Developing models to recognize and transcribe historical scripts

• Training neural networks to read proper names and place‑names in archival documents

• Contributing to cultural heritage digitization and enhancing access to naming data hidden in handwritten paleographic records

A strong computational background is expected, including proficiency in Python, neural network frameworks, and preferably some experience with corpus linguistics or onomastic datasets.

Importance of Named Entities

This project places named entities - personal names, place names, and institutional names - at its core. Accurate transcription of names is notoriously challenging in historical contexts: variant spellings, diacritics, and archaic orthography complicate the recognition process. Successful output hinges on robust models that can handle:

• Ambiguous or low‑resolution name forms

• Rare or archaic surnames and toponyms

• Contextual clues to help disambiguate between homographs

Why It Matters

• For onomastics: The project advances automated extraction of proper names from handwritten manuscripts, feeding valuable data into linguistic, genealogical, and historical research.

• For digital humanities: It supports large‑scale transcription projects in archives and libraries, unlocking naming data embedded in centuries‑old documents.

• For named‑entity research: It demonstrates how machine learning can aid name studies across time - extracting, normalizing, and linking names accurately.

Funding & Timeline

The position is currently advertised by La Rochelle Université in France. As of three weeks ago, the official application window is open. Interested candidates must hold a PhD (completed or near complete) in Computer Science, AI, Computational Linguistics, or a related discipline.

---

✅ Why Onomasticians Should Care

If you're interested in:

• Improving name recognition in historical corpora

• Working at the intersection of computing and name‑studies

• Applying your expertise in anthroponymy or toponymy to real datasets
  this post‑doc offers a unique bridge between AI and onomastic scholarship.

---

🌐 How to Follow & Apply

• Check the EURAXESS portal for full details and deadlines (Job ID: 358464) 

• Prepare a CV highlighting any work on named‑entity recognition, paleography, or computational name processing

• Submit your application via La Rochelle Université’s channel as indicated in the posting

Let me know if you’d like help crafting your motivation letter - including how to emphasize your work on proper names in the project context.