input(ML_Enthusiast) : output(ML_Engineer)

Getting predictions in Jupyter Notebook is sufficient till you are in Campus, but what after getting into Corporate? The model developed once has to reach clients in order to benefit the company. Also, it should be capable of scaling and learning from user generated data as and when required.

Learn about all the necessary Data Science and Engineering skills, responsibilities, and best practices for becoming a (Marketable) Machine Learning Engineer in this session which will walk you through the "Design - Develop - Deploy" cycle.

In a nutshell, this interactive session with group activities, and discussions will nurture an ML Enthusiast turning into an ML Engineer.

Transcript

1. Charmi Chokshi, Shipmnts.com @CharmiChokshi input(ML_Enthusiast) : output(ML_Engineer)
2. Last Session...
3. Last Session...
4. પણ સનેડો હજી બાકી છે!!!
5. Let’s have a look at our smartphones
6. Can you spot any app which uses Machine Learning?
7. Answer the following: - What is the primary functionality / feature of the app? - How it has embedded ML to fulfil it?
8. Let’s figure out ML Pipeline
9. 9 Data acquisition Model Deployment Data Cleaning Feature Engineering Model Validation Model Monitoring Model Selection Model Testing Model Training Hyper parameter tuning
10. Let’s figure out the correct order
11. 11 Data acquisition Model Deployment Data Cleaning Feature Engineering Model Validation Model Monitoring Model Selection Model Testing Model Training Hyper parameter tuning
12. 12 Data acquisition Model Deployment Data Cleaning Feature Engineering Model Validation Model Monitoring Model Selection Model Testing Model Training Hyper parameter tuning
13. 13 Data acquisition Model Deployment Data Cleaning Feature Engineering Model Validation Model Monitoring Model Selection Model Testing Model Training Hyper parameter tuning
14. 14 Data acquisition Model Deployment Data Cleaning Feature Engineering Model Validation Model Monitoring Model Selection Model Testing Model Training Hyper parameter tuning
15. 15 Data acquisition Model Deployment Data Cleaning Feature Engineering Model Validation Model Monitoring Model Selection Model Testing Model Training Hyper parameter tuning
16. 16 Data acquisition Model Deployment Data Cleaning Feature Engineering Model Validation Model Monitoring Model Selection Model Testing Model Training Hyper parameter tuning
17. 17 Data acquisition Model Deployment Data Cleaning Feature Engineering Model Validation Model Monitoring Model Selection Model Testing Model Training Hyper parameter tuning
18. 18 Data acquisition Model Deployment Data Cleaning Feature Engineering Model Validation Model Monitoring Model Selection Model Testing Model Training Hyper parameter tuning
19. 19 Data acquisition Model Deployment Data Cleaning Feature Engineering Model Validation Model Monitoring Model Selection Model Testing Model Training Hyper parameter tuning
20. 20 Data acquisition Model Deployment Data Cleaning Feature Engineering Model Validation Model Monitoring Model Selection Model Testing Model Training Hyper parameter tuning
21. Design - Develop - Deploy
22. Let’s form groups 2 mins
23. What is we are gonna do? • 3 different Problems • 3 different Deployment Pipelines • 3 different Solutions
24. What is we are gonna do? 1. Smile and Pay we know... ◦ How about building a Smile and Enter system for the next #DevFestAhm
25. What is we are gonna do? 2. Prediction of the next word on your phones such as GBoard, a keyboard by
26. What is we are gonna do? 3. Integration of the Object Detection System on Maps
27. Design 5 mins
28. Don’t forget to tweet your design! @GDGAhmedabad @WTMAhmedabad @GDGCloudAhm @CharmiChokshi #DevFest #DevFestAhm
29. Develop 5 mins
30. Deploy 5 mins
31. What would you say at your product launch party? 5 mins
32. Don’t forget to invite us at the party…!! @GDGAhmedabad @WTMAhmedabad @GDGCloudAhm @CharmiChokshi #DevFest #DevFestAhm
33. What we did?
34. This is what an ML Engineer does*
35. Charmi Chokshi, ML Engineer @CharmiChokshi Q&A Comments Suggestions
36. Charmi Chokshi, Shipmnts.com @CharmiChokshi Thank you!