JAYANTH VODNALA

Developed AI-powered software applications using Python and Spring Boot, designing backend systems and RESTful APIs/automation frameworks that improved application performance by 32% and streamlined business workflows. Built ML predictive analytics and risk assessment solutions using Decision Trees, LightGBM, Scikit-learn, and XGBoost; improved forecasting accuracy by up to 34%. Designed generative AI solutions using GPT-4 and Claude for document processing platforms, conversational assistants, and workflow automation, improving operational efficiency by 38%. Optimized cloud-native AI/software on GCP with CI/CD, automated testing, and containerized deployments, accelerating release cycles by 40%. Established microservices, AI inference APIs, and distributed backend services to integrate models into production environments with high performance and scalability.

Improved Naive Bayes classification and K-Means clustering for large-scale dataset analysis, boosting customer segmentation accuracy by 34%. Built deep learning models with TensorFlow for predictive analytics and vision tasks, increasing model performance by 38% and improving production outcomes. Implemented LangChain-based AI applications for LLM integration, contextual retrieval, and prompt orchestration to improve response relevance by 41%. Worked on Speech AI solutions (speech recognition, voice processing, transcription, NLU), improving speech-to-text accuracy by 36%. Created GANs for synthetic data generation and augmentation, increasing dataset diversity by 43% and improving robustness. Used Azure services and PySpark for large-scale data processing, improving big data pipeline efficiency by 47%.

Implemented ANN architectures in PyTorch, tuning multilayer feedforward models to achieve 96% improvement in predictive accuracy across structured/unstructured datasets. Built serverless inference on AWS Lambda with event-driven architecture, reducing execution latency by 94% and improving real-time inference scalability. Created data visualizations with Matplotlib to improve interpretability of analytical results by 93%. Managed PostgreSQL operations by designing normalized schemas and writing SQL queries, improving query efficiency by 95%. Used Seaborn for statistical visualizations (correlation, distributions, categorical analysis) to improve pattern interpretation clarity by 94%. Performed preprocessing with Pandas/NumPy to clean and transform datasets, achieving 97% readiness for modeling.