![Diagnostic
[Interactive Dashboards]
Prescriptive
[Recommendations & Automation]
Predictive
[Machine Learning]
Descriptive
[Reports]
What should
I do?
What will
happen?
Why did it
happen?
What
happened? Insight
@frlazzeri @mathew_jaya@frlazzeri @mathew_jaya](https://image.slidesharecdn.com/aiconflondon2018lazzerimathewfinal-181010111739/85/A-day-in-the-life-of-a-data-scientist-in-an-AI-company-6-320.jpg)
A day in the life of a data scientist in an AI company
- 1. A Day in the Life of a Data Scientist in an AI Company Francesca Lazzeri & Jaya Mathew @frlazzeri @mathew_jaya
- 3. What is AI and Why is it so important? @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya
- 4. Computer Vision Audio Processing Natural Language Processing Knowledge Representation Machine Learning Expert Systems AI Technologies Illustrative Solutions … Virtual Agents Identity Analytics Cognitive Robotics Speech Analytics Recommendation Systems Data Visualization Emerging AI technologies Computer vision and audio processing, for example are able to actively perceive the world around them by acquiring and processing images, sounds and speech. The use of facial recognition at border control kiosks is one practical example of how it can improve productivity. Sense Natural language processing and inference engines can enable AI systems to analyse and understand the information collected. This technology is used to power the language translation feature of search engine results Comprehend An AI system can take action through technologies such as expert systems and inference engines or undertake actions in the physical world. Auto-pilot features and assisted-braking capabilities in cars are examples of this Act What is AI? – To sense, comprehend and act @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya Accenture: Why artificial intelligence is the future of growth, April 2016
- 6. Diagnostic [Interactive Dashboards] Prescriptive [Recommendations & Automation] Predictive [Machine Learning] Descriptive [Reports] What should I do? What will happen? Why did it happen? What happened? Insight @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya
- 7. Systems of Intelligence Engage your customers Empower your employees Optimize your operations Transform your products @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya
- 8. Data, Questions and Metrics @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya
- 9. Data is connected Data is accurate A lot of data Example: Predict whether component X will fail in the next Y days Example: Identifiers at the level you are predicting, relevant data collected & feature engineering using domain knowledge Example: Will be difficult to predict failure accurately with few examples Example: Failures are really failures, human labels on root causes Example: Machine information linkable to usage information @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya
- 10. Business scenario Key decision Data Science question Energy forecasting Should I buy or sell energy contracts? What will be the long/short-term demand for energy in a region? Customer churn Which customers should I prioritize to reduce churn? What is probability of churn within X days for each customer? Personalized marketing What product should I offer first? What is the probability that customer will purchase each product? Product feedback Which service/product needs attention? What is social media sentiment for each service/product? @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya
- 11. Defining Performance Metrics @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya Correlation with the Data Science Metric Establish a Baseline Quantify the Metric Value Improvement Translate into a Quantifiable Business Metric Establish a Qualitative Objective Example: Reduce user churn Example: Reduce the fraction of users with 4-week inactivity Example: Statistically significant A/B test is a clean way. If this is difficult, compare the values of the metric before and after the solution Example: Reduce the fraction of users with 4-week inactivity by 20% Example: Current fraction of users with 4-week inactivity = 60%
- 12. Understanding the ML workflow & the Team Data Science Process @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya
- 13. Understand Business Goals Under- stand Data Discover /Gather Data Ingest Data Collaboration & Version Control Transform Data Monitor/ Maintain Model Create Model Deploy Model Share Results with Business Owners Documentation Respond to changes/lessons Debug, Fix, Enhance, etc. Apps Services Data Engines Build a model Publish a model Consume a model @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya
- 14. aka.ms/TeamDataScience Domain expert Solution Architect Data Scientist Visualization Expert Project Manager Executive Sponsorship – IT & Business Data Engineer Integration Engineer @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya
- 15. Azure AI @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya
- 16. Azure Bot Service Azure Cognitive Services Azure Cognitive Search Azure Databricks Azure Machine Learning Knowledge miningAI apps & agents Machine learning @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya
- 17. Sophisticated pretrained models To simplify solution development Azure Databricks Machine Learning VMs Popular frameworks To build advanced deep learning solutions TensorFlow KerasPytorch Onnx Azure Machine Learning LanguageSpeech … Azure Search Vision On-premises Cloud Edge Productive services To empower data science and development teams Powerful infrastructure To accelerate deep learning Flexible deployment To deploy and manage models on intelligent cloud and edge Cognitive Services @frlazzeri @mathew_jaya
- 18. Azure Bot Service Azure Cognitive Services Azure Cognitive Search Azure Databricks Azure Machine Learning Knowledge miningAI apps & agents Machine learning @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya
- 19. The Team Workspace @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya
- 22. Model Deployment @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya MNIST Dataset http://ufldl.stanford.edu/wiki/index.php/Using_the_MNIST_Dataset
- 26. Scoring file.py Python Environment Azure Kubernetes Service (AKS) or Azure Container Instance (ACI) @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya
- 30. Scoring file.py Python Environment Azure Kubernetes Service (AKS) or Azure Container Instance (ACI) @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya
- 32. Scoring file.py Python Environment Azure Kubernetes Service (AKS) or Azure Container Instance (ACI) @frlazzeri @mathew_jaya@frlazzeri @mathew_jaya
- 33. Azure Machine Learning Services: https://aka.ms/AMLServices Visual Studio Code Tools for AI: https://aka.ms/VSCodeToolsAI Data Science Virtual Machine: https://aka.ms/AzureDSVM https://aka.ms/TeamDataScience http://ufldl.stanford.edu/wiki/index.php/Using_the_MNIST_Dataset @frlazzeri @mathew_jaya
- 34. Francesca Lazzeri & Jaya Mathew @frlazzeri @mathew_jaya Learn more Thank you!