Hassan Soliman –
AI/ML Engineer & Researcher

Hello, I am currently working as a Researcher at DFKI in Berlin. I earned a Master’s degree in CS with an excellent grade at Saarland University in Germany. During my Master’s, I worked at Bosch in Germany, Amazon EU in Luxembourg, and Max-Planck Institute für Informatik in Germany.

INTERESTS

Natural Language Processing (NLP)

Generative AI

Large-scale Language Modelling (LLM)

Conversational AI

Chatbots & Retrieval Augmented Generation

EXPERIENCE

German Research Center for AI (DFKI)

Researcher
Jan 2023 - Present

• Joined the Educational Technology lab, owned two projects technically, and supervised two research assistants.
• Developed a chatbot for a graduate-level course at a German University, that answers students' questions based on the lecture, seminar, and organisational information, achieving 86.7% correct responses, and published two papers.
• Implemented a sub-module in dialogue systems, that adapts the assistant's utterances based on the user's emotional state and demographic properties, benchmarking the results using OpenAI LLMs with two open-source models.
• Supported team members in their projects by developing an Information Extraction system, which extracts words or a span of words from text based on a pre-defined set of keywords, and contributed to project acquisition meetings.

Bosch Center for Artificial Intelligence (BCAI)

Applied Scientist Intern
May 2022 - Aug 2022

• Joined NLP & Semantic Reasoning group, and Utilized my master thesis output in Neural Entity Linking for a real use case in a top project.
• Refactored, tested & documented my code, and worked on training & evaluating the developed ML system on large-scale domain-specific data.
• Model performance is evaluated by comparing with the existing classical system, and achieved a 77% end-to-end recall at the top 3 entities.
Berlin, Germany
Renningen, Germany

Bosch Center for Artificial Intelligence (BCAI)

Master's Thesis Student
Jun 2021 - Jan 2022

• Joined NLP & Semantic Reasoning group, collaborated with research scientists & engineers, and worked on Cross‑domain Neural Entity Linking.
• Utilized a state‐of‐the‐art model to make it easy‐to‐extend‐domain, and experimented which data & information sources are best for fine‐tuning.
• Achieved an Average Precision gain of 9% for the top‑1 entity, and published a paper in RepL4NLP at ACL 2022 venue.
Renningen, Germany

Max Planck Institut for Informatics (MPII)

Research Assistant
Nov 2020 - May 2021

• Joined Database & Information Systems group, and created a prototype for a model that finds diverse peer groups for an entity.
• Implemented a baseline model on the set expansion for an entity that exploits Wiki lists as a source of knowledge.
• Accomplished a 3x faster runtime of the underlying algorithm using an efficient sparse matrix multiplication approach.
Saarbrücken, Germany

Amazon

Software Development Engineer Intern
Aug 2019 - Feb 2020

• Joined Fulfillment Acceleration team, and owned a web application simulation tool by working on it as a Full-stack, using AWS cloud platform.
• Worked in an Agile environment as a system admin to maintain the server hosting team’s tools, and acted as technical support for the team.
• Developed a tool for data visualization of supply chains on maps, and provided workshops for the Business Analysts & Intelligent team members.
Luxembourg, Luxembourg

EDUCATION

Saarland Informatics Campus, Saarland University

Master of Science in Computer Science
GPA: 1.40 / 1.00

Faculty of Engineering,

Alexandria University

Bachelor of Science in Computer & Communication Engineering
GPA: 3.96 / 4.00
Saarbrücken, Germany
Alexandria, Egypt

ACCOMPLISHMENTS

Udacity Foundation

Deep Learning Nanodegree
2018

Alexandria University

First Class Honor Degree
2018
2022
2018

PROJECTS

Scalable Mentoring Support with a LLM Chatbot

Research Project
Designed and implemented a chatbot based on an LLM-based Agent to provide scalable educational support and timely feedback to students of education sciences, demonstrating significant potential of generative AI in education.
Utilized Advanced techniques in Retrieval Augmented Generation (RAG) to enhance chatbot interactions, e.g., Hybrid Ensemble Search and Reranking Mechanism, enabling it to retrieve and analyze course materials effectively.

Using LLMs for Adaptive Dialogue Management

Research Project
• Adapted user-directed utterances using LLMs based on the user's parameters like gender, age, and sentiment, aiming to optimize user satisfaction in conversational AI systems, focusing on healthcare patient-practice interactions.
Evaluated different LLMs and open-source tools for effectiveness in utterance adaptation, in terms of speed, cost-effectiveness, and quality of the generated text based on the adaptation relevancy and adaptation adequacy.
2024
2023

Information Extraction Pipeline in Medical Text

Research Project
• Extracted symptoms from medical text data (prescriptions) in the German language based on a set of doctor-predefined symptoms and their synonyms.
Extracted additionally other symptoms based on a comprehensive ontology for the symptoms, which is provided by the German Ministry for health.
2023

Better Diet to fight COVID-19

Master's Project
• Analysed food consumption from all countries to investigate a relationship between country food culture \& their recovery rate, using Analytics.
• Implemented three Random Forest models and benchmarked the results to report the best one, and displayed figures using Seaborn library.
2021

TF-IDF-based Information Retrieval

Master's Project
• Built a system that takes a query as input and returns a certain number of relevant documents, using various statistics.
• Implemented a second step to retrieve and rank relevant sentences from the relevant documents, achieving a Mean Reciprocal Rank of 0.393.
2020

PUBLICATIONS

THESES

Cross-Domain Neural Entity Linking

Master's Thesis
• Contributed to a single system that enables simultaneous linking of named entities to a general-domain KB (Wikipedia) & a domain-specific KB.
• Utilized Semantic Search by learning a joint vector space for these KBs from different domains using contextual-aware embeddings by BERT.
• Experimented with four different domain-specific KBs, and achieved an increase of up to 20% in Mean Average Precision for the top-10 entities.
• Helped in writing an invention report for my thesis contribution, which is published as a US patent application, and received an Incentive-Prämie.

Egyptian Car License Plate Information Detection

Bachelor's Thesis
• Implemented an application which extracts license information from car images in Egypt, going over the different product life cycle stages.
• Collected datasets for various kinds of car plates in Egypt in various conditions, and applied different Data Transformation (ETL) techniques.
• Utilized pre-trained CNN models for fine-tuning for Object Detection, Localization, Semantic Segmentation, and OCR for the letters & numbers.
2022
2018

PREPRINTS

ArXiv | Effective General-Domain Data Inclusion for Machine Translation by Vanilla Transformers

Jan 2021

Built and trained a Transformer from scratch on the German-English translation task applications of WMT’13.
Utilized a general-domain dataset from IWSLT'16 TED talks to help improve performance of the Transformer model, achieving a 25.8 BLEU score.

ArXiv | Offensive Language Detection & Classification on Twitter

Aug 2019

• Trained a classifier to detect offensive tweets from Twitter using SVM, after performing iterative experiments.
• Achieved a Binary Accuracy of 74% in classifying offensive tweets, and received the highest score among all participant teams.

ArXiv | Data Augmentation using Feature Generation for Volumetric Medical Images

Jun 2019

• Proposed using U-net and ACGAN as a learning framework for feature generation of medical images of two complex types of brain tumors.
• Deployed a classifier pipeline to test & validate the quality of the generated features.

Interested? Let’s get in touch.