TensorFlow Developer Certification: How I passed the exam
 | Rohit Kumar Mahadev |
TensorFlow is an end-to-end open source library for machine learning(ML) from Google. It has a flexible ecosystem of tools that lets developers and researchers build and deploy easily the state-of-the-art ML powered applications.
TensorFlow team offers a certification program to demonstrate and recognize practical machine learning skills through building and training of models to solve deep learning problems
I have been going through ML from past 6 months and set a mini-milestone along my learning journey to take certification exam. Utility of skills beyond certifications is of utmost importance to me. As with everything in life, you gotta be tested once in a while. After all it is a matter of fun to crack open something and quench your thirst of curiosity.
So I made a curriculum to pass the exam based on
- Exam Syllabus: TensorFlow Developer Certification Handbook available for free.
- Book: Deep Learning with Python by Francois Chollet
- Course: DeepLearning.AI TensorFlow Developer Professional Certificate
- Practice: Exercises and labs mentioned in above book and course.
- Environment: Installed free PyCharm Community Edition on my 2014 Macbook Pro(No GPU) to practice and also to take exam. You may also practice the code in the colab provided in course.
| Topic | Book(Chapters) | Online Course |
|---|---|---|
| Basic skills | Chapters 1 - 2 | Introduction to TensorFlow for Artificial Intelligence, Machine Learning, and Deep Learning. Week-1 |
| Build and Train Simple Neural Network Models | Chapters 3 - 4 | Introduction to TensorFlow for Artificial Intelligence, Machine Learning, and Deep Learning Week-2 |
| Image Classification | Chapter 5 | Introduction to TensorFlow for Artificial Intelligence, Machine Learning, and Deep Learning Week 3-4 |
| Natural Language Processing | Chapter 6 | Natural Language Processing in TensorFlow |
| Time Series, Sequence and Prediction | Chapter 6 | Sequences, Time Series and Prediction |
Reading the book slowly to gain intuition, and watching course lessons to understand concepts along with practice code in colab and/or pycharm is good enough to clear the exam. No GPU is necessary unless you want to finish the exam early within stipulated time(5 hours).
Further details about exam environment setup and FAQs are avilable here. You'll get a feedback right away each time you submit the model during the exam.
After passing the exam,
- you are requested to join fellow TensorFlow developers in TensorFlow Certificate Network in the congratulations email.
- A digital certificate will be issued within couple of days.
Here is mine
Using the above curriculum, If I am able to do it,you can do it as well. Best wishes!