Mar
09
Apr
Eventi
Tech Talk DATA REPLY: From Machine Learning Operations (MLOps) to Large Language Models Operations (LLMOps)
Tech Talk DATA REPLY*: From Machine Learning Operations (MLOps) to Large Language Models Operations (LLMOps)
The increasing use of Machine Learning has led to the extension of DevOps principles to the world of data science.
Thus, MLOPs was born to keep models reliable, consistent and efficient in production environments. In recent times, Generative AI has revolutionized the field and opened up possibilities previously not even imagined, particularly in the field of text generation with LLMs. However, these tools work very differently from previous ML models and require an evolution of the MLOPS paradigm to handle their special features: for example, how do you retrain a model that needs thousands of GPUs to be trained? How do you evaluate performance? These and other tasks require a new approach, which is called LLMOPS.
Internship and/or job opportunities offered by the company will also be presented during the event.
The event will be in English.
*Data Reply is the Reply group company offering a broad range of advanced analytics and AI-powered data services. We operate across different industries and business functions, working directly with executive level professionals and Chief Officers enabling them to achieve meaningful outcomes through effective use of data. We support companies in designing and implementing human-centered data products to turn embryonic ideas into enterprise solutions. We aim at handling trade-offs among business requirements and technical constraints, developing holistic Big Data architectures and ensuring cybersecurity and data protection.
The increasing use of Machine Learning has led to the extension of DevOps principles to the world of data science.
Thus, MLOPs was born to keep models reliable, consistent and efficient in production environments. In recent times, Generative AI has revolutionized the field and opened up possibilities previously not even imagined, particularly in the field of text generation with LLMs. However, these tools work very differently from previous ML models and require an evolution of the MLOPS paradigm to handle their special features: for example, how do you retrain a model that needs thousands of GPUs to be trained? How do you evaluate performance? These and other tasks require a new approach, which is called LLMOPS.
Internship and/or job opportunities offered by the company will also be presented during the event.
The event will be in English.
*Data Reply is the Reply group company offering a broad range of advanced analytics and AI-powered data services. We operate across different industries and business functions, working directly with executive level professionals and Chief Officers enabling them to achieve meaningful outcomes through effective use of data. We support companies in designing and implementing human-centered data products to turn embryonic ideas into enterprise solutions. We aim at handling trade-offs among business requirements and technical constraints, developing holistic Big Data architectures and ensuring cybersecurity and data protection.