Siva Satyanarayana Raju Pusapati

• Built agentic AI workflows using LangGraph to coordinate multi-agent tasks and manage context across reasoning steps. Integrated agents with LangChain and RAG for dynamic response generation.
• Leveraged LangSmith for agent telemetry and observability, tracking tool usage and model outputs to fine-tune behavior and improve end-to-end performance of deployed agents.
• Developed production-ready agents using Google ADK, enabling tool calling, calendar scheduling, and decision support workflows. Integrated agents with OpenAI and Flask for full-stack AI applications.
• Developed and deployed AI-driven chatbots using LangChain, RAG, and LLaMA 3, integrating FastAPI and Flask for seamless real-time interactions. Optimized vector storage with ChromaDB, improving query performance by 35%.
• Designed and managed end-to-end ML pipelines, leveraging AWS SageMaker, Azure ML, and MLflow for model training, versioning, and deployment. Built CI/CD workflows with Terraform, Jenkins, and Azure DevOps for automated model deployment and monitoring.
• Built and operationalized MLOps frameworks for AI model performance management, implementing drift detection, bias monitoring, and automated retraining pipelines using Kubernetes, Docker, and cloud-native orchestration tools.
• Developed AI-driven solutions utilizing NLP, deep learning, and LLMs, fine-tuning foundation models on AWS Bedrock and OpenAI GPT. Applied NER, sentiment analysis, and transformer-based architectures for enterprise AI applications.
• Designed scalable AI architectures on AWS, GCP, and Azure, integrating event-driven workflows with Kafka and Spark for large-scale real-time data processing. Automated ETL and data transformation processes using Apache Airflow and Databricks.
• Enhanced AI security and compliance by implementing role-based access control (RBAC), identity management (MS Entra SSO, Azure AD SAML), and integrating security policies with Azure Key Vault and Microsoft Purview. Developed AI-driven cybersecurity measures to identify and mitigate vulnerabilities in AI agent systems.
• Applied computer vision techniques for real-time image classification and object detection using OpenCV, TensorFlow, and PyTorch. Optimized model performance for low-latency AI applications in edge computing environments.
• Conducted research on LLMs, contributing to advancements in generative AI, prompt engineering, and AI-powered automation. Designed cognitive search solutions and vectorized knowledge retrieval systems to enhance decision-making capabilities in enterprise AI solutions.
• Implemented text embeddings to segment and structure documents, enabling efficient storage and retrieval of information from a ChromaDB vector database for accurate, contextually relevant responses.
• Created a responsive frontend application with Django, providing a seamless user experience for interacting with the AI assistant.
• Developed supervised models using LSTM for patient and physician review systems, applying NLP techniques like TF-IDF vectorization and word-to-vector conversion.
• Monitored model performance in production, using logging and alerting systems to detect and address issues promptly.
• Utilized Python for scripting and employed PyTorch and TensorFlow for model development and training.

• Utilized Pandas and NumPy for data cleaning, feature engineering, and normalization, preparing datasets for effective modeling.
• Developed and implemented predictive models using machine learning algorithms such as linear regression, classification, Naive Bayes, K-nearest neighbors, and PCA for comprehensive data analysis.
• Conducted comparative analysis of deep learning models, including LSTM, DRCNN, GRU, ResNet, and RNN, for traffic flow prediction, assessing strengths and weaknesses.
• Utilized performance metrics like Mean Absolute Error (MAE), R-squared, and Root Mean Square Error (RMSE) to evaluate predictive accuracy.
• Preprocessed datasets to create input-output pairs for training and testing, ensuring effective model training and evaluation.
• Designed and implemented predictive models using TensorFlow and PyTorch, experimenting with architectures like CNNs and RNNs for sequential data analysis.
• Developed statistical algorithms in Python, R, and SQL, including Multivariate Regression, Logistic Regression, PCA, and Random Forest models.
• Configured a Python REST API framework with Flask and FastAPI, delivering interactive OpenAPI Standard (Swagger) documentation.
• Conducted customer segmentation using AWS S3, Python, and K-means clustering in Scikit-Learn to enhance marketing strategies.
• Applied advanced NLP techniques to extract actionable insights from unstructured data sources.
• Integrated AI/ML models into production using AWS SageMaker for seamless deployment and scaling.
• Created interactive dashboards in Tableau to provide insights into transaction activities and fraud detection metrics.