User details
Sri Hari Sirisipalli
@sri_hari_sirisipalli_z9dxzpg55
Machine Learning Engineer & Cloud Infrastructure Engineer
Patnam, Andhra Pradesh, India
1Follower0Following0Mutual0Post
Member since: Jan 22, 2025• Predictive Maintenance Development: Led the design and implementation of machine learning models and data pipelines for advanced predictive maintenance, enhancing health monitoring of naval vessel equipment.
• Advanced Fault Detection Algorithms: Engineered and applied machine learning algorithms such as Random Forest, SVM, and k-NN and Deep Learning Modes to improve fault detection accuracy and adaptability across diverse operational conditions.
• Comprehensive ML Pipeline Implementation: Built and refined a full machine learning pipeline from data preprocessing to deployment, providing a robust and scalable foundation for predictive maintenance applications.
Scalable LLM Workflows: Designed dynamic AWS architecture to spin up and terminate EC2 instances on demand, optimizing resource use and cost.
Lambda Health Checks: Automated audits to identify and decommission unused resources, improving performance and cutting costs.
End-to-End AWS with Terraform: Built scalable, maintainable AWS infrastructure including Amplify and monitoring systems.
Custom Terraform GUI: Developed a GUI for easy AWS resource management, enhancing efficiency and accessibility.
Automated Geometry Modeling Pipeline: Developed an end-to-end automated pipeline for wind turbine geometry modeling and simulations in ANSYS APDL using Python, significantly streamlining the integration and processing of design data.
ML-Driven Simulation Analysis: Implemented machine learning pipelines to train and refine models for analyzing simulation outputs, improving the accuracy and precision of performance predictions.
Efficient SAS-to-PySpark Migration Translated complex SAS scripts into optimized PySpark code, improving scalability and performance for large-scale data processing.
Automated Validation Scripts: Developed comprehensive scripts to compare SAS and PySpark outputs with high precision, generating HTML reports that highlight discrepancies to ensure accuracy and validate migration results.
Data Pipeline Dependency Mapping: Applied graph theory to uncover and visualize intricate file dependencies within data pipelines, enabling bottleneck identification and enhancing overall efficiency.
Cross-Framework Model Conversion: Converted deep learning models into multiple formats, facilitating seamless deployment across diverse platforms and frameworks.
Framework-Agnostic Model Integration: Ensured smooth transitions of models from TensorFlow and PyTorch to various deployment environments, optimizing cross-platform compatibility.
Computer Vision Model Optimization: Enhanced the performance of computer vision models by applying advanced techniques such as model compression, quantization, and pruning.
Efficiency without Compromise: Improved model efficiency while maintaining high accuracy, ensuring robust and reliable deployment in resource-constrained environments.
Polkadot-Dojima Integration: Seamlessly integrated the Polkadot ecosystem with the Dojima network chain, enabling enhanced interoperability and functionality.
API Monitoring Dashboards: Designed and implemented Prometheus-Grafana dashboards to monitor APIs in real-time, improving system observability and performance tracking