Dicoding: Incorporating AI to Maximize Bangkit Students’ Learning Experiences

Nurturing future tech talents since 2020, Bangkit has trained over 20,000 participants with industry-relevant skills: Machine Learning, Cloud Computing, and Mobile Development (Android). This career readiness program is initiated by Google, supported by GoTo, Tokopedia, and Traveloka, and under the auspices of the Kampus Merdeka program by the Ministry of Education, Culture, Research, and Technology (MoECRT).

Google appointed Dicoding to be responsible for Bangkit program management and curriculum development, ensuring equitable distribution of tech education across Indonesia. It resonates with Dicoding’s vision: to be a leading tech education platform that promotes broader access to digital literacy for all.

In 2024, with Google Indonesia, Dicoding incorporated Artificial Intelligence (AI) into the Bangkit curriculum. Google stated on its Google AI site that it wants to make AI helpful for everyone and is committed to improving the lives of as many people as possible. Knowing that AI is now an integral part of tech advancement, we aspire for Bangkit graduates to be AI-empowered digital talents. Besides, many industries rely on this technology to augment their business today.

Utilizing AI to Prioritize the Students’ Learning Experience

Since 2015, Dicoding has emerged as a prominent tech learning platform due to our submission review system. This feature sets Dicoding apart because our priority is improving our students' skills. Therefore, to prepare Bangkit graduates for the AI-driven workforce with proven tech skills, Dicoding develops an AI curriculum instilled in Bangkit’s 3 (three) paths and incorporates AI into our submission review system to maximize the student’s learning experience.

Using Google Vision AI, we developed an AI-assisted submission grader. We built this system for 2 (two) weeks, continued research and development to improve this technology for 3 (three) months, and deployed the feature in one of our courses: Menjadi Google Cloud Engineer (Becoming a Google Cloud Engineer), a mandatory class for Bangkit participants opting for Cloud Computing path.
“Our AI-assisted submission grader can accurately scan images uploaded by students and convert them into text (OCR),” Dimas Maulana, Curriculum Developer Lead at Dicoding Indonesia, said.
At the beginning of this feature’s development, Dicoding utilized self-hosted Tesseract OCR. However, this technology's accuracy limitations hindered its ability to process submissions effectively. Finally, Dicoding switched to using Vision AI from the Google Cloud Platform to achieve more accurate and reliable results for production use.

AI-Assisted Submission Grader: the Faster, the Better

By implementing an AI-assisted submission grader integrated with Google Cloud Vision AI, Dicoding could optimize the review process of all first submissions in the Menjadi Google Cloud Engineer (Becoming a Google Cloud Engineer) course with an 81% confidence level. Moreover, Vision AI service helped Dicoding reduce up to 99% of the cost. Furthermore, the review time for assignments has been reduced by 10 (ten) times while maintaining quality and achieving a participant satisfaction rating of 4.9 out of 5.

As a result, Dicoding reaped positive feedback from the students:
“The review process was swift,” Adam Faishal Wijan said.
“My codes were reviewed thoroughly!” Aan Putra Bungsu added.
Besides the Menjadi Google Cloud Engineer (Becoming a Google Cloud Engineer) course, Dicoding is currently developing another AI-assisted submission grader for the Belajar Membuat Front-End Web untuk Pemula (How to Create a Front-End Web for Beginner) course, supported by Gemini AI technology.

“Since the user’s learning experience is Dicoding’s top priority, we are committed to continuously innovating with AI that could effectively address the challenges the students face throughout their learning journey. We hope this AI-assisted submission grader will also be able to be implemented in other Dicoding classes,” Dimas Maulana closed his statement.


Popular posts from this blog

The Journey of a Young Woman Who Dared to Dream Big

Land the Dream Job as a Software Engineer at a Multinational Tech Company

My Physical Condition Is Not a Barrier to My Ambition