ARAVIND GARAPATI

• Developed Generative AI-based customer interaction model, integrating LLMs for automating telecom support and enhancing customer engagement.
• Algorithm Optimization: Integrated LightGBM (LGBM) algorithm, which is known for its efficiency and performance, especially on large datasets with high-dimensional features
• Conducted thorough hyperparameter tuning and applied advanced techniques like early stopping to optimize model's predictive performance.

Principal Accomplishment: Successfully integrated GenAI-driven chatbots into customer service workflows, reducing manual resolution time by 40%.

• Developed and implemented computer vision algorithms for object detection, recognition, and tracking, resulting in 30% improvement in accuracy and 35% reduction in false positives
• Designed and implemented Yolo-SSD Deep learning models for image detection and classification .

Principal Accomplishment:

• Created and Integrated production machine learning pipelines and models with a daily processing capacity of 100,000 photos.

• Carried out exploratory data analysis for customer in supply chain management, which led to discovery of process inefficiencies.
• Developed Deep learning based LSTM forecasting model for retail client resulting in 20% operational cost savings .
• Forecasting models were put into practice for inventory management at client store locations, which improved logistics supply system and produced more accurate project

Principal Accomplishment: This model improved region's total average repair time KPI by 28%.

• NLP, BERT Transformers, and torch framework were used to develop Personal Identification Information (PII) Detection Machine learning model.
• Used BERT to create new entity recognizer model based on transformers. Pre-trained Architecture & Model
  included new data science engine into the offering to address the shortcomings of the previous Spacy/regex based solution.

Principal accomplishment:
By successfully training the Data Science Engine with BERT Mechanism instead of Spacy/Regex code, we were able to achieve Critical PII Entity Tagging accuracy of 80% from 52%.

• Developed proof of concept for content-based, cooperative filtering and user-driven recommendations for junior library clients in US schools.
• To suggest products to clients, a machine learning recommendation system model was built and integrated with the customer's e-commerce website.

• Predictive models were created in order to estimate service demand, spot possible disruptions, and allocate resources as efficiently as possible.
• Worked together with cross-functional teams to create data-driven service improvement plans and identify key performance indicators (KPIs).
• Automated data collection, processing, and visualization tasks using Python, SQL, and Tableau, reducing manual effort by 30%.
• Carried out root cause analysis on service incidents and put corrective measures in place to stop them from happening again, which led to a 20% decrease in the number of occurrences.

Key accomplishment: Spearheaded efforts to enhance service delivery procedures and apply ITIL best practices, which raised customer satisfaction ratings by 15%.

• Performed routine patches and critical software updates for BSS Nodes and Switches.
• Software installation, setup, integration, validation, and troubleshooting of current solutions on client devices.
• Third-level technical support for technical issues pertaining to BSS operations.