Soundarya Hiremath

• Company Overview: DISH - U.S. Telecom Industry.
• Streamlined two use cases (incident and RCA) resolutions for NOC engineers with a centralized dashboard.
• Contributed to the Network Incident Management team, with an emphasis on data management from multiple sources, and worked with embeddings while choosing appropriate AI models, applying AI techniques for NOC dashboards.
• Enhanced troubleshooting efficiency by unifying various data sources in a single dashboard.
• Configured PostgreSQL and Milvus vector databases for efficient data retrieval through vector search and summarization techniques.
• Enhanced resolution efficiency for network-related issues by 95%.
• Specialized in LLM-based response retrieval.
• Developed AI pipelines to optimize chatbots functions.

Gained over 1 year of experience in Python and AI/Gen AI. I showcased my expertise in AI by working with clients and internal initiatives.

AI Projects:

• MACHINE LEARNING WING (AXIOS): Utilized TensorFlow, Pandas, Scikit-learn, and CV2 frameworks for implementing highly accurate CNN-based image classification.
• HAND SIGN LANGUAGE CLASSIFICATION: Engineered a precise classification system for hand sign languages using various libraries and frameworks, like TF, Torch, Matplotlib, and OpenCV. .
• SENTIMENT ANALYSIS ON AMAZON PRODUCT REVIEWS: Designed a robust sentiment analysis solution with 90% accuracy, transforming e-commerce insights and optimizing customer satisfaction strategies.

GENERATIVE AI Projects:

• Voice assistant app with OpenAI's GPT-3 and IBM Watson : Integrated IBM Watson speech-to-text and OpenAI GPT-3 models for enhanced user interaction.
• Summarization of private data with generative AI and RAG: Developed a chatbot with Llama 2 LLM and RAG techniques, achieving 88% accuracy in summarizing PDF documents.

• Company Overview: National Grid - a multinational electricity and gas utility company.
• Directed the development and deployment of customized applications for essential Maximo improvements for National Grid US, resulting in a 99% boost in user satisfaction.
• Solved call bounding process issues through strategic problem-solving.
• Managed complete software development lifecycle from requirements gathering to deployment.