INTERNSHIP OFFER 
Natural Language Processing / Information Extraction / Digital Humanities
(4 months, Bac+5: M2 or Engineering degree)

Internship Topic 
Named Entity Recognition and Entity Linking in 
Early-Modern Prints and Manuscripts

Keywords: 
handwritten text recognition (HTR), optical character recognition (OCR), 
named entity recognition (NER), entity linking (EL)

Duration: 4 months (start: March or April 2026)

Application deadline: March 12, 2026

Compensation: legal French internship allowance 
(approximately ¤650/month)

Host Laboratory:
BdTln Team, LIFAT (EA 6300), Université de Tours 
64 avenue Jean Portalis, 37200 Tours, France

Context

Early-modern printed books and manuscripts present significant
challenges for automatic text processing: non-standardized spelling, 
language variation, complex document layouts (paratexts, marginalia), 
and transcription noise introduced by OCR and HTR systems. These 
factors greatly reduce the performance of downstream natural language 
processing (NLP) tasks, especially named entity recognition (NER) and 
entity linking (EL). Overcoming these issues requires specialized 
methods, including normalization strategies for historical spelling, 
domain-specific annotation guidelines, and training or fine-tuning 
models on historical corpora. In collaboration with the ERC project 
PRIMA - Manuscripts in the Age of Print (Grant Agreement No. 101142242), 
hosted by the Centre d'Études Supérieures de la Renaissance (CESR, UMR
7323, Université de Tours), this internship focuses on developing
and evaluating NER and EL techniques applied to early-modern prints,
manuscripts, and bibliographic catalogs.


Tasks

The intern will be responsible for:

-   Annotating named entities in early-modern texts to create a 
    reference (gold-standard) corpus
-   Developing, applying, and assessing NER and EL approaches on 
    OCR/HTR outputs from historical documents
-   Analyzing results in relation to transcription noise and document
    features, and measuring their impact on entity extraction and 
    linking performance


Internship Framework

This internship is financed by La Région Centre-Val de Loire (France) 
and involves collaboration between the LIFAT and CESR labs at 
Université de Tours, as well as partners specializing in manuscript 
studies and NLP. Regular interactions with the PRIMA research team are 
planned.

The intern will be hosted at Université de Tours in the LIFAT
laboratory (Tours). Travel between Tours and Blois (with
reimbursement) is planned for team meetings, where the intern will
participate. Possibility exists for continuation as a PhD project,
depending on results and funding.

Desired Profile / Required Skills

-   Final-year Master's student or engineering student in Computer
    Science, Computational Linguistics, or related fields
-   Strong programming skills in Python and experience with NLP 
    libraries and model fine-tuning (e.g., NLTK, spaCy, Hugging Face)
-   Familiarity with large language models (LLMs), including practical
    experience with APIs
-   Interest in early-modern texts and digital humanities (prior
    experience is a plus but not required)
-   Ability to document methods accurately and conduct rigorous
    experimental assessments

Application

Applications should include a CV and cover letter, and must be sent by
March 12, 2026, to the following addresses:

Carlos-Emiliano González-Gallardo, LIFAT/CESR, Université de Tours
gonzalezgallardo@univ-tours.fr 

Cyril de Runz, LIFAT/IUT Blois, Université de Tours
cyril.derunz@univ-tours.fr 

References

https://hal.science/hal-05248289v1
https://univ-tours.hal.science/hal-04662000/
https://ceur-ws.org/Vol-3180/paper-84.pdf