Thomas John

Traffic Partner Blacklisting Project:

• Led online shop performance enhancement project, focusing on improving traffic quality.
• Developed system using Python and PySpark for comprehensive traffic and profit analysis.
• Optimized budget allocation using linear programming, enabling efficient blacklist creation.
• Established performance-based blacklisting criteria, thus improving the overall gross profit and traffic quality.

Reporting Tool Evaluation Project:

• Led a comprehensive 3-month project to evaluate and recommend the optimal reporting tool for the organization, focusing on future integration with cloud-based solutions and enhancing the current on-premise big data infrastructure.
• Conducted an in-depth analysis of Looker, Power BI, and the existing on-premise cube-based reporting tool.
• Evaluated tools based on criteria such as web-based development, version control, collaboration features, self-service capabilities for business users, and performance.
• Collaborated with cross-functional teams to gather requirements and understand current and future reporting needs.
• Developed and executed test cases to compare performance, usability, and integration capabilities of each tool.
• Presented findings and recommendations to senior management, highlighting the pros and cons of each solution, and aligning them with our strategic goals.

• Conducted financial data analysis using Python, Hive, and SQL for accurate forecasting.
• Utilized Apache Spark for large-scale data processing, delivering timely financial insights.
• Followed agile methodologies and managed projects with Jira, improving collaboration and efficiency.
• Migrated data science projects from SAS to Python, enhancing flexibility and scalability.
• Applied Dataiku DSS for efficient data engineering, including preparation and transformation.
• Trained team members on Dataiku DSS and Apache Spark.
• Engaged in collaborative coding and version control with GitHub.
• Presented financial forecasts using Tableau, aiding data-driven decision-making for stakeholders.

Semiconductor Wafer Defect Signature Analysis:

Developed a computer vision model to analyze scanned images of semiconductor wafers, detecting defect locations and shapes (e.g., circles, lines, blobs).

AI-based approach for MRO Optimization and Similar Parts Detection:

Designed an AI-driven solution to identify and eliminate duplicate spare parts from a client's inventory of over one million items, enhancing inventory management and approval rates.

Log analysis-based failure monitoring in Hadoop cluster

• Built a monitoring system to monitor the resource usage of the cluster by monitoring the applications running on YARN.
• Designed and built statistical analysis models using Apache Spark ML

• A network controller used to improve the network traffic monitoring and analysis
• Got excellent feedback from customers after implementing a rule based classification algorithm to automatically prioritize the ACL rules to redirect the traffic to the corresponding monitoring services.

• The Linux Foundation's OPEN-Orchestrator Project & A SDN based service aware Network Controller and management system

• Web-based network management solution that offers configuration, provisioning and troubleshooting for a wide range of technologies
• The system managed multiple network devices, and provided management for services across different network elements

• Web-based software application that helps to automate the daily operations for the service provisioning of access NEs, such as DSLAMs