Chithira Raj Email & Phone Number

Mississauga, ON, CA

Mississauga, Ontario, Canada

• Client: Citibank
• Enhanced a machine learning model utilizing FastText to automate data lineage tracing, refining the mapping accuracy of destination fields to their source fields across relational database tables.
• Employed reinforcement learning with Q-learning to reward correct source and destination data sources and penalize incorrect ones, leading to a 90% improvement in data lineage accuracy.
• Analyzed debt data break details to assess the accuracy of the data lineage mappings, identifying inconsistencies and improving data traceability across systems.

Toronto, Ontario, Canada

• Extended the Robocop Python tool by integrating controls that calculate customer rewards and check for exceptions, utilizing standardized SQL script execution and enhanced data quality checks.
• Implemented validation processes to ensure accuracy in control scripts and developed reporting features for comprehensive data monitoring.
• Standardized control execution for customer data monitoring by automating processes with scheduled jobs and alert systems for timely notifications to stakeholders.

Canada

• Created a web scraping script with BeautifulSoup to extract information for summarizing articles using an LLM model. Fine-tuned the LLM Flan-T5 model for article summarization.
• Applied ARIMA, Prophet and HWES for analyzing time series data to predict future sales trends.

Bengaluru, Karnataka, India

• Led the development of unsupervised hierarchical clustering model with 80% score to detect ESXi host failures by log analysis.
• Optimized log analysis processes by implementing scalable data pipelines using PySpark for data extraction, leading to improved model performance.
• Developed anomaly detection models using K-Means and Isolation Forest, to identify abnormal log patterns with 70% performance.
• Implemented a VPT based nearest neighbor search algorithm to enhance host failure prediction accuracy by 80%.
• Evaluated model performance, initiated statistical testing, and optimized model parameters to align with business objectives and regulatory requirements.
• Prepared executive level Tableau / Power BI dashboards and reports to communicate data insights, model performance and future strategies.

Bengaluru, Karnataka, India

• Implemented a Gradient Boost regression model using Azure ML to enhance invoice processing of Microsoft’s financial transactions, achieving an accuracy of 85%, expediting invoice approvals and payments.
• Created Power BI dashboards to communicate invoice processing model insights with the stakeholders.
• Designed a real-time predictive analytical tool and dashboard using HWES time series forecast model with a 90% accuracy in predicting future productivity.
• Collaborated with stakeholders to understand business requirements, evaluated model performance, ran statistical tests, optimized parameters, and informed strategic decisions, boosting business performance.

Bengaluru, Karnataka, India

• Developed a XGBoost Classification model to predict the invoice payment outcomes, with 80% accuracy.
• Incorporated NLP techniques to implement sentiment analysis and topic classification model on vendor communication data, improving the accuracy by 30%.
• Predicted support ticket inflow using time series algorithms, achieving 90% accuracy.

Chennai, Tamil Nadu, India

• Proposed a facial expression training platform for autism with an accuracy rate of 80%, learning and applying Convolutional Neural Network (CNN) and Haar Cascade Classifier techniques.
• Acquired proficiency in Data Structures, Java, Python, and SQL through dedicated training.

Master'S Degree, Data Science

Part-Time / Online Course designed for working professionals.

Bachelor Of Technology - Btech, Computer Science