CHANDRAJYOTHI HARI

FCS Ingestion application is responsible for processing claim files which are from different sources. Claims go through basic validation process under FCS Ingestion application and will be sent to further downstream applications for claims approval.

Responsibilities:

• Designed and implemented data warehouse concepts and data models, ensuring data accuracy and consistency.
• Executed SQL queries for data extraction and manipulation, maintaining data integrity and optimal performance.
• Implemented Apache Airflow for pipeline setup, using advanced concepts like DAGs and Xcom for efficient workflow orchestration.
• Increased team's efficiency by 30% by learning and implementing new technology with minimal assistance.
• Maintained up-to-date records by consistently updating Jira tickets, ensuring clear communication and project progress tracking.

Express Ordering Application: Account validation, address validation and email notification once after a successful order placement. Microservices hold a vast number of mS applications which are used as APIs in different applications. All applications are targeted to migrate to Azure

Responsibilities:

• Orchestrated migration of Express Ordering application to Azure cloud, optimizing code structure and enhancing performance.
• Increased efficiency up to 50% of Scalability, maintenance and upgrades.
• Partnered with cross-functional application teams to resolve bugs, implement code changes, and manage Jira tickets, ensuring streamlined operations and delivery.
• Demonstrated expertise in Python, Django, Azure, Kubernetes, Docker, and SQL, facilitating efficient migration and integration processes.

Text search tool: It is a cognitive AI-SME assistant tool, built as an Information Retrieval (IR) system to support IT teams on their natural language knowledge queries from existing Knowledge sources. The corpus for this system includes millions of MS Word documents and tickets from different IT applications. This tool provides an easy-to-use web interface to query and retrieve desired information.

Responsibilities:

• Enhanced IT team productivity by integrating millions of documents and tickets into AI-SME assistant tool.
• This tool has started with increase in productivity by 20% and improved till 45%
• Aggregated and standardized critical data from multiple AT&T systems, improving information retrieval and project understanding.
• Improved data quality for machine learning models by removing stop-words, lemmatization of tokens, and utilizing advanced NLP techniques.
• Developed and refined Sequence-to-Sequence deep learning model with LSTM, achieving 30% increase in precision.
• Launched Flask-based API for seamless integration of trained model, enhancing customer satisfaction through feedback loop with user feedback API using Cloudant DB.

Cognitive Metrics Tool: Proactive quality metrics management system which visualizes, predicts quality as per organization specific aspects and features. Dynamic multi-dimension outputs regarding quality data are presented for easy track-back and project management.

Responsibilities:

• Conducted requirement analysis and collaborated with stakeholders, including Quality Management team, to design and develop asynchronous RESTful microservices using Spring Boot, DB2, and Java 8, achieving average response time of 30 milliseconds.
• Designed and implemented over 60 microservices APIs leveraging Java 8 features within Spring Boot framework for various metric calculations and managed DB2 database operations.
• Assisted in developing Power BI reports, conducting data analysis, and supporting data validation and quality assurance processes.
• Utilized Hibernate with JavaBeans for transactions, implemented CI/CD pipelines using Jenkins, and employed factory design pattern in utility development.

Predicting Order Fall out / Task Fall out in Ordering Applications. Involved in building machine learning models using Scikit-learn Pipeline using Keras deployed in AT&T CMLP runtime for predicting task level / order level fall outs based on data from End-to-End Flow Management System (EFMS) & Global Integrated Order Management system (GIOM) in the service delivery area.

Responsibilities:

• Conducted in-depth interviews with Subject Matter Experts (SMEs) for comprehensive requirement analysis.
• Gathered and prepared data conducted data wrangling, and established data pipelines.
• Engaged in feature selection, label encoding, and split of training and testing datasets for model development.
• Evaluated model performance using Scikit-learn's classification report.
• Utilized Scikit learn pipeline, Python pandas.

Automating Pareto Bucketing of Order fall out tickets for CBS team (BVOIP, ADI and AVPN order fall out tickets). Intention of this solution is to classify tickets created for Order fall outs using machine learning models and prepare Pareto

Responsibilities:

• Increased SME efficiency by automating Pareto bucketing of CBS order fallout tickets
• This tool has replaced all manual efforts with 90% automation.
• Developed machine learning model (CNN) to classify order fallout tickets, reducing SME's effort in ticket lifecycle.
• Managed data collection and preparation processes for ticket classification using Selenium and Beautiful Soup.
• Implemented data pipeline setup, ensuring streamlined feature selection and model training.
• Achieved higher model accuracy by evaluating with confusion metrics, thereby improving ticket classification.

IMA-Access: IMA access allows customers to place service requests through IMA-GUI or IMA-XML. IMA-BPL is C++ application which controls all business functionalities and channels to other applications for further processing.

Responsibilities:

• Developed business and application logic for IMA-BPL, streamlining voice and data service delivery.
• Led development of IMA-BPL, C++ application, enhancing telecom service request processing and improving specificity and relevance of telecom service projects.
• Developed business and application logic, prepared Low-Level Design (LLD) and interface specifications, and automated routine tasks, resulting in 25% reduction in development errors and 15% increase in operational efficiency.
• Accurately estimated efforts for change requests, executed user-based testing, and documented results to enhance product quality and optimize project timelines.