Machine Learning Engineering Fundamentals
Topic: Machine Learning Engineering Fundamentals
Presenter: denny wang
Additional Resources:
MLE Fundamentals
Coach Ken LinkedIn:
https://commitway.com/linkedin
WeChat QRCodes
| | | 职场提升俱乐部 |
Trends in MLE role
Is MLE a specialized or general role?
Similar to software engineering, MLE role will broaden to include different parts of ML projects
Is MLE different from SDE?
In the long run, they will have a very large overlap
How to transition into ML projects?
Knowledge
Hands-on
Find opportunity
Will the large models replace the small models?
Small models will continue to exist
Medical image recognition
Is better at explaining the results
Large models will replace small models in some situations.
Natural language processing
Content generation
There are some possibility for vertical industry with lots of data to replace small model with large models
How to overcome limitations of large model, such as privacy and safety?
Overlay a large model with a smaller model
Or train a new model
Most of the ML related jobs cover the full lifecycle of the ML development
Need to understand how to convert a business problem into a ML problem
Need to be good at integration
Trend:
From building tool to using tool
Access latest models:
AWS jump starter
Paper with code
Huggingface
Kaggle
These are great for step 1 in learning AI/ML
1 hr
3 directions for MLE work
Model
Pipeline
Monitoring MLOps
AI/ML Tools for automation
Integration with existing pipeline
Are there higher requirements to work with ML Model?
Not difficult if there is opportunity
Reading paper is useful