Sri Charan Bodduna
Data Scientist - Generative AI
Passionate about exploring data and finding insights using Large Language Models and data science tools. Capable of building Machine Learning applications to solve business problems. Building highly efficient models with best solutions in the industry.
What I Do
Ethical AI Development
Follows ethical standards of data and AI, ensuring responsible development practices.
Generative AI Solutions
Building Generative AI solutions for business use cases & automation of manual tasks.
Predictive Modeling
Developing flawless models to predict outcomes and improve business performance.
Model Optimization
Optimizing models to produce results in the most efficient way possible.
Scalable Solutions
Scaling ML designs to be data-driven and cost-effective with industry standards.
Business Intelligence
Building Business Intelligence solutions to help businesses make data-driven decisions.
Projects
Peterbilt Cost Optimization
Developed highly efficient Machine Learning ensemble models to optimize maintenance costs, leading to better decisioning between high and low-cost truck configurations with 88% accuracy. Secured top 5 in the hackathon.
Research Copilot – AI Research Assistant
ChatGPT-based Generative AI driven Research assistant that helps professors and students understand research papers in 5 minutes with interactive QnA capabilities.
Preventive Care Attendance Classification
Developed a solution for classification of 1.5M+ patients who do not visit for preventive care using Gradient Boosting techniques, validating the model with 75% AUC and 70% accuracy. Intensive feature engineering and critical data understanding. Deeper insights using machine learning algorithms with stronger focus on ROC-AUC characteristics.
JobBuddy.ai
Developed an LLM application that can act as friend during job search. The application is capable of resulting key words matching score between resume and job description. Other capabilities include description of resume with reference to job description provided, how to improve skills, motivations during job hunting, and drafting LinkedIn recommendation message.
Handwritten digits classification using TensorFlow
Implemented a deep learning AI model to classify handwritten numeric digits on MNIST dataset using TensorFlow. Achieved 98.8% accuracy with two layered Convolution Neural Network (CNN).
Dallas Crime Analysis Reporting using Tableau
Developed a eye catching data visualization on city of Dallas police arrests. I have found many insights about arrests patterns and frequency of arrests in the city. These outcomes can be used by city managers to view and strengthen the law and order of the city.
Work Experience
University of North Texas
- Developed an ethical data science model assessment game in collaboration research project with 16% improvement in understanding the ethical implications of AI deployments.
- Implemented test ML models and visualizations using Python, and JavaScript (d3js), leading to a 14% increase in performance accuracy assessment & facilitated creation of an interactive game.
Vodafone Intelligent Solutions (VOIS)
- Delivered $29k in savings by uncovering balance sheet disparities through TensorFlow deep learning models and time- series analysis for profits forecasting, enabling precise calculation of statutory damages and penalties.
- Supervised 5 servers and 600+ workflows under Vodafone Group, assisting the company in adapting new ETL pipelines. Administered users, workflows & collections on servers.
- Written and Optimized SQL queries using CTEs and Stored Procedures reducing execution time by 24%. 12% increase in profit growth margin.
Infosys Ltd.
- Designed a Data Quality application using Alteryx and Python (spaCy) for HSBC, that assesses completeness, consistency, validity, and conformity, resulting in 92% time savings in data management and pipeline processing by automating the app.
- Optimized the application measuring data quality for 10 million rows and 500+ columns using Alteryx and its server capabilities, by reducing execution time by 96%, leading to significant savings in business hours and enhanced operational efficiency.
Education