10-11.06.2024 : Introduction to MLOps

 

Event:  Introduction to MLOps

Date: 10-11 June 2024

Location: Building G22A - Room 122

Event Summary:

The Academic Club at OVGU, FIN recently hosted a two-part workshop titled "Introduction to MLOps" on 10th and 11th of June. The workshop, was designed to provide participants with a foundational understanding of MLOps, focusing on best practices, the machine learning lifecycle, and practical tools for managing and deploying machine learning models.

 

Day 1: Exploring MLOps Best Practices and ML Lifecycle

During the first session, participants were introduced to the key principles of MLOps, emphasizing best practices that ensure effective collaboration between data scientists, ML engineers, and operations teams. The session provided an overview of the machine learning lifecycle, covering the stages involved in developing, deploying, and maintaining ML models. The session included a hands-on demonstration with Weights & Biases, a powerful tool for tracking experiments, visualizing model performance, and optimizing hyperparameters. Participants had the opportunity to explore Weights & Biases, gaining practical experience in how to integrate these tools into their own machine learning workflows.

 

Day 2: Model Deployment Techniques

The second session focused on model deployment, an essential aspect of MLOps. Participants were introduced to FastAPI, a modern web framework for building APIs with Python, and Gradio, a user-friendly interface library for deploying ML models with minimal code. Additionally, the session covered how to leverage HuggingFace's tools for model deployment and management. Through guided exercises, participants gained hands-on experience in deploying machine learning models using these tools. By the end of the session, they had successfully deployed a simple ML model, gaining confidence in their ability to transition from model development to deployment.

 

Conclusion

The workshop was well-received, with active participation and engagement from all attendees. Over the course of the two sessions, participants gained valuable insights into MLOps, equipping them with the knowledge and tools to effectively manage the machine learning lifecycle and deploy models efficiently. We look forward to organizing more such events in the future to continue fostering learning and innovation in the field of machine learning.