About Nana Agyei-Kena

  • Academic Level Master’s Degree
  • Industry Development
  • Viewed 22

About me

• Research and development focused on data-driven solutions using Machine Learning and Deep Learning to enable autonomous processes in the field of related projects in NLP and AIoT e.g. Risk Management in connected industries, Analytics on Distributed Data Storage, Knowledge graphs for Data-Driven Solution, Neural Machine Translation, Sentiment Analysis, Chatbots for Data-Driven solutions, etc.



  • 2021 - 2021
    Ferdinand-Steinbeis Institute, Stuttgart

    NLP/Machine Learning Engineer - Working Student

    NLP Tasks
    • Implemented a Knowledge Graph (KG) for AI Data-Driven Solutions using
    Spacy, Transformers, Stanford parser and Neo4j graph database.
    • Implemented a Custom Spacy NER & Relation Extraction for entity named
    recognition on a large text corpus for AI Data-Driven Solutions
    • Implemented a Custom Fine-tuned BERT Relation Extraction (RE) for
    relation extraction on a large data-driven text corpus.
    • Implemented a Fine-tuned Stanford Parser for Entity Extraction (EE) for a
    large AI Data-Driven text corpus
    • Successfully lead a small group and implemented a Content Management
    Software (CMS) for question and answering on AI Data Driven Solutions
    ML Tasks
    • Risk Management: Implemented an Early Fire Detection and Prevention
    (EFDP) using ML algorithms and ensemble learning for a Saw-mill Co.
    achieving an accuracy over 80% and reducing further environmental risk by
    • Made presentations and meaningful reports on Exploratory Data Analysis on
    Temperature time-series data for Risk management.
    • Anomaly Detection: Implemented an Unsupervised Anomaly detection for
    Temperature time-series data using an ensemble of models (Isolation
    forest, SVM, GMM and KNN)
    • Implemented a Service Design on the EFDP for sustainable solutions and
    optimal experiences for both the customer and the Institute
    • Implemented a Conceptual Model to represent the system and workflow of
    the EFDP
    • Managed the IIC resource Patterns website for the institute


Machine Learning
Deep Learning
Html CSS Javascript Flask



Honors & awards

  • 2020

    Design Patterns Based on Deep Learning analyzing Distributed Data

    Data silos in many system landscapes complicate the creation of comprehensive information. Distributed Ledger Technology enables trust between assets via a distributed, secure and immutable storage of transactions. Deep Learning realizes intelligence to make decisions, conduct them and analyze their results based on the gathered data. In order to counteract the limitations of current system landscapes, an integrative implication of both Distributed Ledger Technology and Deep Learning is needed. Many considerations arise during the design of information systems integrating the named technologies. Transparency over the training and deployment of various Deep Learning methods on a distributed data landscape needs to be achieved. Using both a literature review and a qualitative research approach, this paper describes the development of design patterns and their selection criteria with different dimensions taken into consideration. The evaluation phase comprises of semi-structured interviews with experts from different disciplines. The result of this paper guides stakeholders in the selection of a suitable technical solution.


Leave Your Review

  • Overall Rating 0