Senior Data Scientist and Computer Vision Expert

Project: Cidaas-ID-Validator

2018 - Present

• Directed the development and deployment of scalable microservices to operationalize machine learning models, enhancing system efficiency and responsiveness.
• Innovated an OCR (Optical Character Recognition) model tailored for high-precision information extraction from various identity documents, significantly improving the accuracy of data retrieval processes.
• Engineered a sophisticated security check framework for identity verification, integrating Computer Vision and Probabilistic Graphical Models to detect fraudulent documents effectively.
• Pioneered the design and development of a Human-Computer Interaction (HCI) application leveraging eye-tracking methodologies (ninoxipy) for real-time liveness detection, setting new standards in identity verification technology.
• Played a pivotal role in the conceptualization and implementation of an MLOps framework, facilitating seamless model integration, monitoring, and management, thereby ensuring continuous improvement and operational excellence of deployed machine learning solutions.

Expert Data Scientist

Project: Cidaas-FDS (Fraud Detection System)

2020 - Present

• Architected an ML-based cybersecurity solution, focusing on the integration of advanced fraud detection mechanisms within the system’s core to bolster security measures.
• Engineered RESTful APIs to seamlessly serve fraud detection models, enabling real-time fraud analysis and immediate response to security threats.
• Instrumental in the development and maintenance of a CI/CD pipeline, ensuring the agile deployment and rigorous testing of models, which contributed to a streamlined development process.
• Innovated a Smart Multi-Factor Authentication (MFA) system utilizing a Fuzzy Inference System (FIS), which enhanced user verification processes by adapting to varying risk levels.
• Conducted Causal Inference analysis to identify the root causes of fraudulent activities, enabling the implementation of targeted preventive measures.
• Achieved a breakthrough in fraud detection accuracy, reaching a very high accuracy rate of 97%, significantly reducing the risk and impact of fraudulent activities on the platform.

Senior Data Scientist

Project: Bosch-EBR (Experience-based Repair)

2020 - Present

• Pioneered the development of an NLP-based vehicle identification system, employing Named Entity Recognition (NER) and Graph Theory to accurately identify and classify vehicle components.
• Architected and implemented a RESTful API, facilitating the efficient delivery of the vehicle identification model to stakeholders, enhancing user accessibility and system integration.
• Designed and developed a robust stream processing framework using Faust for EBR microservices, optimizing data handling processes and improving the system’s scalability and resilience.
• Contributed the creation of an LLM-based Concept Identification Service, leveraging large language models to mine car part failure discussions in forums, and algorithmically identifying approved solutions to address these failures. This innovative service significantly improved the efficiency of diagnosing and resolving vehicle issues by directly linking reported problems with validated solutions.

Lead Data Scientist

Project: Porsche-CCD (Corner Case Detection)

2018 - 2019

• Engineered a state-of-the-art spatio-temporal model that leverages internal sensory data and the geo-location of Porsche cars to detect corner cases, significantly enhancing the predictive capabilities and safety measures.
• Spearheaded the design and development of an interactive UI for monitoring Porsche cars, incorporating ReactJS to create a responsive and user-friendly interface that enables real-time visualization of data and identification of plausible corner cases.
• Developed a high-performance data streaming system utilizing MQTT and Tornado, ensuring efficient data transmission and processing capabilities for real-time monitoring and analysis.
• Implemented advanced data visualization techniques using Plotly and Bokeh, enabling complex data insights to be comprehensively understood through interactive graphs and charts, thus facilitating informed decision-making processes.

Computer Vision Expert and Data Scientist

Project: Chinese Painting Seal Assessment

2018 - 2019

• Pioneered the development of a cutting-edge deep learning model using state-of-the-art Convolutional Neural Networks (CNNs) in TensorFlow to accurately identify seals of Chinese artists, enabling the authentication of artworks with high precision.
• Innovated a recurrent neural network model designed to assess the originality of artists’ seals, effectively distinguishing genuine artifacts from forgeries. This model leveraged the nuanced patterns and historical data of seal imprints, contributing significantly to the preservation of cultural heritage and the art market’s integrity.

Researcher and Developer

• Spearheaded the innovative modeling of gaze movements within the Immersive Quantified Learning lab (iQl), advancing the field of Human-Document Interaction through groundbreaking research that bridges the gap between digital learning environments and natural user engagement.
• Led the design and implementation of a cutting-edge Human-Computer Interaction system utilizing eye-tracking technology for the AICASys Project, markedly enhancing the accessibility of digital interfaces for individuals with disabilities, thereby democratizing technology use.
• Developed a novel generative probabilistic model for synthesizing eye-movement patterns, which stands as a significant contribution to understanding and predicting user interactions in digital environments, paving the way for more intuitive and adaptive systems.
• Contributed to academia as a Teaching Assistant for the Data Mining Course at TU Kaiserslautern, enriching the educational experience of students with practical insights and fostering a deep under- standing of critical data analysis techniques.
• Supervised Master and Bachelor dissertations in the fields of AI and HCI at TU Kaiserslautern, guiding students through their academic journeys with mentorship that emphasizes innovation, critical thinking, and real-world application of theoretical knowledge.

Full-stack Developer

• Designed a comprehensive personality assessment project for Mapna Group, successfully developing and deploying a robust evaluation system for over 20,000 employees. This initiative significantly contributed to enhancing workforce management and organizational development through tailored psychological insights.
• Demonstrated full-stack development skills by designing and developing a dynamic portal for organizastional psychology assessments. Utilized PHP for server-side logic and Python for data analysis and processing, showcasing a versatile proficiency in both front-end and back-end technologies. This portal facilitated seamless, efficient assessments, highlighting my ability to deliver comprehensive digital solutions from concept to deployment.

Programming Languages

Python, R, C++, JavaScript, PHP, Go

Libraries

Numpy, Scipy, Pandas, scikit-learn, PyCaret, scikit-fuzzy, scikit-image, OpenCV, dlib, PyTorch, Causal Inference, Tensorflow, Keras, Statsmodels, Django, Flask, FastAPI, Sphinx

Databases

MySQL, PostgreSQL, MongoDB, Neo4j, Redis, Elasticsearch

Visualization Tools

matplotlib, seaborn, bokeh, plotly, ggplot2, Tableau

IDEs

PyCharm, Visual Studio Code, Sublime-text, RStudio

MLOps Tools

GitLab CI/CD, Docker, Kubernetes, MLFlow, Apache Kafka, DVC, CML

Operating Systems

Linux (Fedora and ubuntu), MacOS, Windows