Bangkit 2021 Experience — Capstone Project
Earlier this year, I got the opportunity to join a Machine Learning Bootcamp at Bangkit Academy by Google, Gojek, Tokopedia, and Traveloka. I learned a lot on this journey, and I want to share several things I experienced there. More about my Capstone Project.
Curriculum — ML
As the Machine Learning Path Participant, our curriculum is highly related to Python, Math, and Tensorflow. We also have many soft skills classes to improve our chance of getting jobs and growing into professionals. Unlike last year's Bangkit program, this year’s program is designed to be fully online, with most of the classes are provided by Coursera and Dicoding platforms and the rest are Google Meet Session with Instructors from various backgrounds. Bangkit Program applies self-paced learning so you guys can work at your own pace and finish before the deadlines.
The Machine Learning Path Curriculum begins with finishing two courses of Google IT Support and Google IT Automation with Python, though It’s good for a newbie, I didn’t quite like the course. I felt the course is too lengthy and irrelevant to Machine Learning. Next was a course on Mathematics for Machine Learning, this course gives us some necessary knowledge on matrices and PCA, about the math behind Machine Learning algorithm. This topic is really heavy and hard to follow, but I think It’s really valuable for a researcher who prioritizes the math and algorithm behind ML.
After a month, we finally got to our main topic, Deep Learning Tensorflow Development. This course teaches us how to create a deep learning model by doing projects, and this is the most fun course in this program. There are 4 main topics, which are CNN, NLP, Time Series. Our next course is focused on the common technique used to develop Deep Learning Model to keep efficiency and goals achieved. Some technique was particularly important to avoid bias and false results, that only looks good in Training but particularly bad in deployment.
Lastly, our final course is focused on how to deploy our trained model to various deployment situations. For example, deploying as a browser-based model or also on a device like Raspberry Pi.
Curriculum — Softskills
During our program, we are provided with various soft skills material and meeting sessions. We have been through topics like personal development, career growth, critical thinking, idea creation, and business mindset. Our meeting sessions are led by various instructors from the industry, the sessions last 1–2 hours and often involve a discussion and participation from the students.
Lastly, a Capstone Project
This is what every participant has been waiting for, the capstone project begins in late April with around ~2900 participants left. In this year's program, we were given a choice to either make our own team of 6 (consist of 2 members from each learning path) or be randomly selected by the committee. Each team needs to create a project to solve real-life problems that combine the Machine Learning, Android, and Cloud Computing Path into their project.
Keep in mind that handling ~2900 participants is no easy task, many teams failed the requirement or have a missing participant. I got lucky to form my team early, giving me opportunities to choose my teammates and having more solidarity early on. For others, some have their Bangkit experience ruined due to missing teammates or got shuffled from their initial team to another team.
At the end of the Bangkit program, the committee is going to choose 40 teams out of 500 teams into the final, and 15 chosen teams will get funding from the industry. This means there are chances to kickstart your project into a real product and turn it into a company or a startup.
My takeaway
I think Bangkit is a good program to kickstart your tech career, the program allows you to start a real-life project with a mentor and teammates, which is a really valuable experience to me. The self-paced aspect of Bangkit means you had to discipline yourself to your assignment, if you were passionate and determined on learning, then Bangkit is for you. There will be many mentors from the industry that you can talk to, and many others like you to be your rivals/Nakama.
While some aspects could be improved for Bangkit, I think the Bangkit Team has done a great job managing 3000 participants into the initial project goals. I’m looking forward to next year’s Bangkit and see how it will go.
Tags: Bangkit, Capstone Bangkit, Bangkit Experience, Bangkit Blog, Google Bangkit, Pengalaman Bangkit, Bangkit2021.