PyLinea - Data Lineage Tool (Product Manager, Scrum Master, Architect, Core Developer):
Led the conception, development, and implementation of PyLinea, a groundbreaking Data Lineage Tool. As Product Manager, Scrum Master, and Core Developer, I devised the entire architectural framework, utilizing abstract syntax tree notations and a robust graph database for end-to-end lineage within the Data Lake ecosystem.
Key Responsibilities:
- Led the conceptualization, design, and development of PyLinea, defining the strategic vision and objectives of the tool. Conducted thorough assessments of business requirements and engaged with stakeholders, including Chief Data Officers, to align PyLinea with organizational goals.
- Innovated and formulated the entire architectural framework of PyLinea, ensuring scalability, flexibility, and adaptability to diverse data environments.
- Embraced an agile methodology as the Scrum Master, fostering a collaborative and iterative development process.
- Organized and led scrum ceremonies, ensuring efficient communication within the development team, resulting in a 40% increase in project delivery efficiency.
- Acted as the core developer, leading the actual implementation of PyLinea's features and functionalities. Utilized abstract syntax tree notations for objects to capture intricate data relationships, establishing a clear and concise representation of data flow within the ecosystem.
- Leveraged a robust graph database to store and manage the complex interconnections, enabling seamless navigation of data lineage. Expanded PyLinea ecosystem by introducing complementary products, such as a log parser and a GDPR/sensitive data monitor.
- The log parser facilitated efficient parsing and interpretation of data logs, contributing to improved data governance and compliance.
- The GDPR/sensitive data monitor addressed privacy concerns by identifying and monitoring sensitive data, ensuring adherence to regulatory requirements.
The success of PyLinea extends beyond lineage, creating an integrated ecosystem that significantly enhances data governance and compliance.
Employee Health Service using AI / Safety Oculus (Product Manager, Scrum Master, Architect, Core Developer):
Served as the driving force behind the innovative Employee Health Service using AI, known as Safety Oculus. As the Product Manager, Scrum Master, Architect, and Core Developer, I orchestrated the entire lifecycle of the project, from conceptualization to implementation.
Key Responsibilities:
- Conceptualized, designed, and developed Safety Oculus—an AI system ensuring worker safety by verifying proper PPE usage. Aligned the product strategically with organizational goals, translating scientific insights into practical AI use cases.
- Led architectural design for Safety Oculus, ensuring scalability and efficiency. Implemented Visual Computing techniques (YOLO, Faster R-CNN) for enhanced PPE detection and improved safety compliance.
- Took on the role of Scrum Master, steering agile development to maximize collaboration. Acted as the primary point of contact, fostering communication between key stakeholders, including CIOs, Engineering Site Managers, and IT teams.
- Engaged with Engineering Site Managers to bridge the gap between IT and Engineering teams, ensuring seamless integration and alignment with operational needs. Interacted directly with CIOs to understand their strategic goals and incorporated feedback to enhance the product's effectiveness.
- Translated key scientific findings into AI-driven business use cases, solving complex problems through the implementation of Safety Oculus. Achieved a dynamic Video AI analytic alert system, ensuring real-time safety monitoring and intervention.
- Recognized with the Deliver with Focus Award by GE Power CIO for outstanding contributions to the successful implementation of Safety Oculus.
Serial Number Detection - Deep Learning based object detection process – YOLOv5, Mask-RCNN to detect serial numbers embossed on industrial parts. Driving the development of object detection pipeline and streamlining it into production in AWS and IOT ecosystem in collaboration with GE Research
Tax Type Classification - ML and RPA based utility which classifies the type of transaction based on details of the transaction from the ledger. Designed and developed the multiclass classification model using Categorical Boosting.