Working Student in NLP

Fraunhofer-Gesellschaft · Darmstadt

WerkstudentGerman needed: not said

Still open on 2 Nov

You apply on the company's own website.

Key facts

Implement natural language processing approaches
Extract features using LLMs, embeddings and traditional feature engineering
Implement style change detection algorithms
Detect AI-generated text and identify which LLM generated it
Improve existing authorship analysis methods for attribution and verification
Implement and apply various machine learning methods and concepts
One-class and binary classification using LLMs, such as zero-shot and few-shot learning and foundation model fine-tuning; deep learning approaches, such as CNNs, Transformers and Siamese neural networks; and traditional approaches, such as SVM, Random Forest, Isolation Forest and kNN
Apply ensemble methods, including stacking, voting, weighting and calibration
Evaluate machine learning models using standard metrics, including ROC, AUC and confusion matrices
Develop user interfaces and web apps using fastHTML, Streamlit, Gradio, Angular and Svelte
Contribute to publicly funded projects and projects commissioned directly by industry partners

Very good academic results with an average German grade of 2.2 or better
Strong Python and frontend development skills are essential and will be tested in the interview.
Strong knowledge of machine learning and deep learning
Familiarity with different neural network architectures, including Transformers, MLPs and 1D CNNs
Familiarity with basic terms and concepts such as classification, hyperparameter optimization, fine-tuning and model evaluation
An advantage: ability to independently implement methods and procedures from scientific publications
An advantage: knowledge and experience in cybersecurity
Willingness to take on new challenges
Strong analytical thinking

The work takes place on site at Fraunhofer SIT in Darmstadt. Remote work is not possible.
Good pay: €18.06 per hour for bachelor's students and €18.78 per hour for master's students
Flexible working hours that fit well around your studies
An inspiring working environment with modern infrastructure
The chance to gain practical experience and make valuable contacts in research
Opportunity to complete a bachelor's or master's thesis later
Monthly working hours of 40 to 80 hours, as agreed

Implement natural language processing approaches
Extract features using LLMs, embeddings and traditional feature engineering
Implement style change detection algorithms
Detect AI-generated text and identify which LLM generated it
Improve existing authorship analysis methods for attribution and verification
Implement and apply various machine learning methods and concepts
One-class and binary classification using LLMs, such as zero-shot and few-shot learning and foundation model fine-tuning; deep learning approaches, such as CNNs, Transformers and Siamese neural networks; and traditional approaches, such as SVM, Random Forest, Isolation Forest and kNN
Apply ensemble methods, including stacking, voting, weighting and calibration
Evaluate machine learning models using standard metrics, including ROC, AUC and confusion matrices
Develop user interfaces and web apps using fastHTML, Streamlit, Gradio, Angular and Svelte
Contribute to publicly funded projects and projects commissioned directly by industry partners

Very good academic results with an average German grade of 2.2 or better
Strong Python and frontend development skills are essential and will be tested in the interview.
Strong knowledge of machine learning and deep learning
Familiarity with different neural network architectures, including Transformers, MLPs and 1D CNNs
Familiarity with basic terms and concepts such as classification, hyperparameter optimization, fine-tuning and model evaluation
An advantage: ability to independently implement methods and procedures from scientific publications
An advantage: knowledge and experience in cybersecurity
Willingness to take on new challenges
Strong analytical thinking

The work takes place on site at Fraunhofer SIT in Darmstadt. Remote work is not possible.
Good pay: €18.06 per hour for bachelor's students and €18.78 per hour for master's students
Flexible working hours that fit well around your studies
An inspiring working environment with modern infrastructure
The chance to gain practical experience and make valuable contacts in research
Opportunity to complete a bachelor's or master's thesis later
Monthly working hours of 40 to 80 hours, as agreed