Double Duty: Surviving Bangkit Academy & Landing My First Tech Internship

Tags: Bangkit Academy, Machine Learning, Python, Internship, Web Scraping, College Life

2021: The Peak Performance Year

Hi, it is Tama again. To be honest, this year has been pretty rough but incredibly rich in experiences—much more than I expected. So many things have happened since January, and I cannot wait to see what shows up in the next half. I originally drafted this article around June, and I am just continuing it now after so much more has happened. I truly feel 2021 is my peak performance year in college. I know I will miss this adrenaline rush at some point in the future, so I am writing it down here to make sure I never forget it. Let’s start!

The Kampus Merdeka "Loophole"

This all started around mid-2020 while I was leading the Google Developer Student Club. I remembered hearing about a collaborative program between the Indonesian Ministry of Higher Education and tech giants like Google, Gojek, Tokopedia, and Traveloka. It was part of the Kampus Merdeka initiative (specifically the Independent Study program) initiated by Nadiem Makarim. The goal was to teach practical, industry-standard tech skills to prepare students for their careers. The previous iteration had 300 participants focusing on Machine Learning, and I knew I had to take my shot.

When registrations opened in October 2020 for the 2021 cohort, I immediately told my academic counselor, Mr. Aris. The new program offered three distinct learning paths: Machine Learning, Android Development, and Cloud Computing. The best part? It was an 18-week (6-month) program worth up to 20 university credits (equivalent to a full semester).

Since this was a brand-new program and the campus hadn't established strict conversion rules yet, I decided to exploit this "loophole." My strategy was simple: register, get accepted, and let Mr. Aris handle the complex approval strategy later, hahaha! I chose the Machine Learning path because I had been focusing on it for the past couple of years, and it aligned perfectly with the credits I needed to convert. Strategy at its finest to legally skip campus classes!

Speedrunning Coursera

By January 18, 2021, I received my official acceptance letter, and the onboarding started in February.

Bangkit wasn't just about technical skills; it heavily incorporated business and soft skills. For the ML path, we relied heavily on Coursera. I was already familiar with the platform thanks to the JuaraGCP program (where I previously earned rewards for learning Google Cloud Platform). Normally, a Coursera specialization takes 4–6 months to complete. In Bangkit, we had to finish 4–5 of them in 3–4 months. As a cohort, our running joke was that we had to "speedrun" these courses!

On top of the self-paced learning, we had Instructor-Led Training (ILT) sessions 1–2 times a week. These were live, English-taught online classes led by industry experts covering both technical and soft skills. Bangkit really forced me to be independent and aggressively manage my time, focus, energy, and learning pace.

The Juggling Act: Landing an Internship

Right in the middle of this Bangkit chaos, around March 2021, a fellow GDSC Lead named Timur approached me. His software house in Jogja was looking for a Python developer with experience scraping and extracting data from social media platforms like Twitter, Facebook, and Instagram. Thanks to a lecturer who had previously introduced me to the ropes of web scraping, I knew exactly how to do this.

I updated my CV with my recent Bangkit lessons, listed my projects, and landed an interview. They asked me exactly how I would acquire the social media data. I replied: "Are we doing this the official way or the unofficial way?" They asked to hear both.

I explained that the official way (using their official API) takes time to set up and requires strict compliance. You get the data safely, but the volume is limited unless you buy premium API services. The unofficial way (scraping via web traffic and cookies) allows you to extract as much data as you want, but you risk triggering bot-detection systems and getting your IP or account blocked.

I got the job. I waited a week until the confirmation letter arrived before telling Mr. Aris. That was another calculated move—if I had told him beforehand, he might have told me to decline it since I was already juggling a heavy Bangkit workload. Thankfully, he believed I could manage both.

"Fail Fast" and the Capstone Reality Check

For the final 1.5 months of Bangkit, we had to build a capstone project and present it to a panel of judges, aided by industry Project Advisors.

My team built RoadQue, a system designed to detect potholes and road cracks using a mobile phone. The architecture involved a machine learning model hosted on a server; a mobile app would take a photo of the road, send it to the server, and the model would return an assessment. We were given $200 USD in cloud credits.

Unfortunately, reality hit hard. One team member forgot to pause the testing server, draining our credits almost instantly. We also lacked sufficient data and resources to train the model properly. Furthermore, the team dynamics were chaotic: we had one Cloud Engineer, me as the sole ML Engineer, and two App Developers who only showed up in the final weeks. It was a mess, but I pushed to get as much done as possible. You can't always control the team, but you can control your output. (Repo: roadque-bangkit21)

During one of the main presentations earlier in the program, Mas Yoza (a prominent figure we always saw on Discord and announcements) showed a slide with two words: "Fail Fast."

It started as an inside joke among the participants, but we quickly took it seriously because people were dropping out left and right. The point is, do not be afraid to try new things; if you must fail, fail as quickly as possible early in the process so you can immediately evaluate, learn, bounce back, and fix those mistakes before wasting a lot of time and money.

In a fast-paced environment like Bangkit, this became our coping mechanism for the deadlines, the coding errors, and the relentless exams.

Graduation, Merch, and Certifications

After the capstone project wrapped up, the waiting game began. I received my graduation confirmation around June 25, 2021, and we officially graduated on July 15, 2021.

One of the best perks of joining Bangkit was definitely the merchandise! Throughout the program, students earned points for class participation, attendance, and task completions. I racked up enough points to claim a light jacket, a mug, and some stickers.

Another massive perk was a voucher to take a professional certification exam (distributed on a first-come, first-served basis). The Cloud Computing cohort could take the Associate Cloud Engineer exam, the Android cohort got the Associate Android Developer exam, and for the Machine Learning cohort, we were offered the TensorFlow Developer Certificate. I obviously claimed the TensorFlow voucher and immediately started preparing on the Dicoding platform.

Looking Ahead

Now it is August, about a month since I graduated from that "hell training ground," and I actually miss all the memes we shared.

As for the internship? After Bangkit ended, my initial 3-month contract (which was supposed to end in July) was extended. My superiors wanted me to stay on the project team and dive even deeper into that complex API integration. But that technical deep-dive is a story for another article.

Keep going!