Profilbild von MohammadZahurul Islam CONSULTANT, MACHINE LEARNING EXPERT, SENIOR MACHINE LEARNING ENGINEER aus Berlin

Mohammad Zahurul Islam

verfügbar

Letztes Update: 18.03.2024

CONSULTANT, MACHINE LEARNING EXPERT, SENIOR MACHINE LEARNING ENGINEER

Abschluss: PhD in Computer Science
Stunden-/Tagessatz: anzeigen
Sprachkenntnisse: deutsch (gut) | englisch (gut)

Dateianlagen

islam-01-2024_180324.pdf

Skills

Natural Language Processing (NLP), Machine Learning (ML), Deep Learning, LLMs, Python, Java, AWS, GCP, Git, SQL, NoSQL, Vector Database, Data Science

Projekthistorie

07/2022 - bis jetzt
AI Lead
Liesenfeld Research Institute GmbH, Hamburg (Internet und Informationstechnologie, < 10 Mitarbeiter)

  • Led a team to develop a multilingual data pipeline for automatic conversation transcription, preprocessing, and implicit motive extraction. Extraction accuracy is more than 80%.
  • Headed the development of a tool for identifying mental health problems from speech and texts.

03/2024 - 03/2024
Lead AI Consultant
Millionways Inc., New York (Internet und Informationstechnologie, 10-50 Mitarbeiter)

  • Spearheading the development of implicit motivation identification from telephone conversations and texts.
  • Pioneering the development of a fraud identification tool from speeches and texts.

01/2019 - 05/2022
Professor
NORDAKADEMIE Hochschule der Wirtschaft, Elmshorn (Internet und Informationstechnologie, 50-250 Mitarbeiter)

  • Instructed courses in Python and Data Science courses.
  • Conducted research on Computer-Aided Psychographic Text Analysis.
  • Developed multiple ML solutions analyzing implicit human characteristics, collaborating with researchers from different universities.

10/2020 - 12/2021
AI Consultant
20Flow7 GmbH, Hamburg (Internet und Informationstechnologie, < 10 Mitarbeiter)

  • Delivered an AI model for identifying motives and personality traits from motivation letters.
  • Coordinated with cross-functional teams and stakeholders.

06/2018 - 05/2019
MACHINE LEARNING EXPERT
WAFM WIRTSCHAFTSAKADEMIE GMBH; Wedemark-Mellendorf

- Research on automatic implicit personality analysis.
- Buil tools that extract personality, motives, and sensitivity from the texts in different languages.

Wedemark-Mellendorf, Germany

08/2017 - 05/2018
SENIOR MACHINE LEARNING ENGINEER
SCAN.UP AG (A SISTER CONCERN OF THE PAWLIK GMBH)

* Build a text-categorization system that predicts humans' implicit motives (16 categories) with an accuracy of
more than 80%.
* Also building categorization systems for English, Dutch, Spanish, French and Italian texts from the psychological
domain.
* Responsibility includes planning, design, and implement the categorization systems as micro-services.
* The AWS has used to train machine learning algorithms.

Berlin, Germany

08/2015 - 05/2017
HEAD OF DATA INTELLIGENCE
KEY4TALENT GMBH (PREVIOUSLY KNOWN AS MASSINESCHEFFER GMBH)

* Used Discriminative Multinominal Naive Bayes algorithm for OMT categorization that improves the accuracy by
5%.
* Built the automatic OMT categorizer for English, French, Spanish, and Italian by translating German OMT text
using Google translator. Overall accuracy on the test set from respective languages is more than 60%.
* Had to lead a research project to categorize OMT categories from unrestricted natural language texts.


Berlin, Germany

02/2015 - 07/2015
R&D SOFTWARE ENGINEER
MASSINESCHEFFER GMBH

* Automatize the manual OMT (Operand Multi-motive Test) process using Naive Bays Multi-nominal Algorithm.
* The solution outperforms the manual categorization process by 6% for categorizing OMT texts into 16
categories.



* Implement the Automatic OMT solution as a REST API.

Berlin, Germany

02/2011 - 12/2014
RESEARCH STAFF (WISSENSCHAFTLICHER MITARBEITER)
TEXT TECHNOLOGY LAB, GOETHE UNIVERSITY

* Developed a multilingual source and translation detection tool using Java and Support Vector Machine (SVM). In
domain accuracy 82%.
* Developed a multilingual readability classification tool using similar technology. In domain accuracy more than
80%.
* Proposed several Information-theoretic features for text readability classification and source and translation
identification.

Frankfurt am Main, Germany

10/2009 - 07/2010
GRADUATE RESEARCHER
DEPARTMENT OF COMPUTER SCIENCE, UNIVERSITY OF PISA

* Contributed to the first version of the Italian to English Translation system.
* Participated in the design of dependency phrase-based machine translation.

Pisa, Italy

Reisebereitschaft

Weltweit verfügbar
Profilbild von MohammadZahurul Islam CONSULTANT, MACHINE LEARNING EXPERT, SENIOR MACHINE LEARNING ENGINEER aus Berlin CONSULTANT, MACHINE LEARNING EXPERT, SENIOR MACHINE LEARNING ENGINEER
Registrieren