MLMU Online #15: Key factors of successful MLOps

The talks will be live on YouTube.

Multiple speakers will shows us how they different technologies to successfully implement MLOps with variety of organizations and they will also share the most important and tricky parts of this process.

1. talk: DataOps: an indispensable component of MLOps (Wioletta Stobieniecka, SAS)
   Abstract: ModelOps as one of the leading trends attracting huge interest in entire data science community, software vendors and companies embedding analytics into their everyday processes.
   My presentation is aimed at presenting SAS approach to ModelOps based on best practice of model development and deployment resulting from years of experience. However, I would like to go further extending this idea on the indispensable component of the process – data and thus, the second part will be dedicated to DataOps concept and the idea of Feature Store.
   Bio: Data Scientist with over 6 years of experience, an open-mindedness and a deep passion for solving real-world problems using advanced statistical and machine learning tools. Expert in statistical learning methods, both in theory and practice. Analyst with extensive experience especially in antifraud analytical solutions with major focus on building predictive models in industries like banking, insurance or telco. For almost two years engaged in computer vision related projects applying both classical and deep learning techniques. Broad exposure to both open source and commercial software (Python, SAS, R, and others). Advocate for ModelOps & DataOps. She has recently deep dived into analytics for healthcare.
   https://www.linkedin.com/in/wioletta-stobieniecka-705b0365/

2. talk: Experiment tracking in MLOps with MLFlow (Michal Marusan, Microsoft)
   Abstract: In this short talk we show the importance and means of experiment tracking using MLFlow - crucial part of MLOps model development lifecycle - in the modern cloud ML environment. How to develop your models with governance and traceability in your hands to boost productivity and increase reusability.
   Bio: I am Cloud Solution Architect for Data & AI at Microsoft. My passion is enabling customer to do more with their data through AI & Machine Learning using Cloud technologies. Before that I worked as a Data Scientist at various projects in Telco, Finance and Manufacturing verticals. I have been focusing on Big Data & ML & AI for last 10+ years.
   https://www.linkedin.com/in/mma123/

3. talk: MLOps on AI Platform (Dorian Hodorogea, Google)
   Abstract: The discussion will cover the Unified AI approach in Google Cloud as well as best practices on how AI/ML initiatives should be covered within organizations so that data scientist engage the right teams and drive the ML process organization wide. At the end of the presentation there will he a real life example of a customer journey: options, mistakes, changes and evolution of their platform.
   BIO: Dorian Hodorogea is an "analytics" mind with a curious soul, working as a Customer Engineer for Smart Analytics at Google. He is based in Bucharest and covers customers around CEE helping them find innovative solutions with the help of technology. In the past 12 years, he has covered different industries as well as technologies in the Analytics/AI/ML world while working for Google, SAS Institute and IBM.
   https://www.linkedin.com/in/dorianhodorogea/

Language: English

Machine Learning Meetups (MLMU) is an independent platform for people interested in Machine Learning, Information Retrieval, Natural Language Processing, Computer Vision, Pattern Recognition, Data Journalism, Artificial Intelligence, Agent Systems and all the related topics. MLMU is a regular community meeting usually consisting of a talk, a discussion and subsequent networking. Except for Prague, MLMU also spread to Brno, Bratislava and Košice.

http://www.mlmu.cz/