Dr. Rafiqul Haque

Cognitus Paris - France

Speaker 3

DBLP: https://dblp.org/pers/hd/h/Haque:Rafiqul

Rafiqul Haque is the Chief Technology Officer and Co-founder at Cognitus – a research, innovation, and development center based in Paris, France. At Cognitus, he leads the data science teams and research activities. He manages the Big Data projects for the industrial clients of Cognitus. Also, he is the project manager of the Shyfte4.0 – a European Union Erasmus+ project aims at building an industry driven academic program for building next-generation experts.
Over the years, Rafiqul worked in different research organizations and companies in different countries. He was experimental track manager and researcher a Chaire d’excellence project at laboratoire d’informatique en image et systèmes d’information (LIRIS). He was a research fellow at Laboratoire Hubert Curien (Telecom Saint-Etienne, France), Université de Versailles Saint-Quentin-en-Yveline (France), Governance Risk and Compliance Competence Centre (University College Cork, Ireland), and European Research Institute of Service Science (Tilburg University, the Netherlands). He worked for SAP AG, Germany as a research assistant and developer.
Rafiqul has worked in various projects funded by the European Union, Enterprise Ireland, Science Foundation Ireland, and ANR (France). Over the years, he contributed in various research areas including Big Data Engineering and Analytics, Scalable and Distributed Computing, Service Oriented Computing (SOC) and Cloud Computing. He led the development of several innovative methods for data engineering, data analysis, and service-oriented computing. He has published several conference and journal articles available in IEEE Xplore, ACM Digital Library, and Springer database. Rafiqul serves as a program committee member of several conferences sponsored by IEEE and ACM.

Talk 3: Fallacies and Critical Success Factors of Enterprise Data Science Project: An Industry Perspective

Enterprise data science is far beyond the fantasized hype of Big Data articulated in various media. In industrial reality, there is no place for a buzzword. Therefore, enterprises demand a great deal of value from data science and AI products and services which eventually ensure the return of investment. They expect data science and AI to lead to digital transformation that enables organizations to adapt to the fourth industrial revolution and lead to achieve business excellence and sustainable growth of the enterprise. 
Manage the real-world value-added industrial data science projects require strenuous effort as various challenges and pitfalls can be fostered from different fronts. For instance, very often the preferred technologies increase effort during deploying and maintaining data science project artifacts and assets. Defining the right strategies for adopting disruptive Big Data technology is a critical challenge because it may lead to a massive digital transformation within and across the enterprise solution landscape. Reproducibility of performance in terms of accuracy is outstandingly difficult to avail; for example, any model, data, and feature set versioning are mystic if an enterprise wants to track overtime which model trained on what data gave what kind of performance metrics?
The “Silo Mentality” within the working culture of enterprises is a common pitfall in a data science project. It is a mindset, present when certain departments or sectors do not wish to share information with others in the same company. This for certain is a substantial barrier for enterprise data science projects. Having a silo mentality leads to loss of productivity, efficiency, reduced morale and eventually projects end up getting scrapped. Furthermore, often people doing data science projects thinking, ‘these are software engineering methodologies and principles — we are well above this’. However, in practice there is none; in other words, there is no standard methodology and architecture. This can sink a huge hole in complex machine learning, deep learning, and AI model, leading to yet another failed proof-of-concept that could not make it to production.
In this talk, I will focus on some very real challenges plaguing the industry concerning executing data science projects. Also, I will discuss the driving factor that leads to the success of enterprise data science projects. I will discuss some strategic guidelines help to gain some useful perspective towards the effective execution of data science projects. The guidelines are based on distilled personal experiences and by looking at industry trends and talking to industry experts.