As I've told you before, since 2022 I've been pursuing my master's degree. I took it while I was working full-time at my previous employer. Luckily I worked hybrid, so I just need to go to the office at Bandung about twice a month. I can say for almost two years, I experienced long burnout due to academic and career pressure. You can imagine I have to work from morning to evening, then continue to study at night.
The biggest challenge was when I had to finish my master thesis. I didn't want to take the machine learning-related topic in the first place. But again, sometimes things change. I got an advisor lecturer who has expertise in machine learning and cyber security and he forced me to take the machine learning implementation for the risk management domain which made me have to think hard.
Another thing that made me struggle enough is not only the thesis itself. I have to prepare myself to publish my thesis in a paper publication as a graduation requirement. This is why for several months after I finished my thesis exam, I still struggled to finish my research publication.
Here's my story from writing the thesis proposal until the publication finished.
When I first embarked on my master’s journey, I had no idea that my path would lead me into the intricate world of machine learning and risk management. After securing an advisor who was enthusiastic about these fields, I found myself both excited and overwhelmed. The advisor’s push to explore machine learning for risk management was a thrilling yet daunting challenge. However, what initially seemed like a labyrinth of confusion soon became a navigable path, thanks to the invaluable support I received from unexpected quarters.
One of the most fortuitous turns in my journey was connecting with supportive seniors on Twitter. Their guidance was a beacon of hope during my early days of confusion. In particular, I remember a pivotal moment in January 2023 when I had a one-on-one session with Mas Aria Ghora and Eric (algonacci). Their insights were a game-changer for me. They helped me see that the seeds of my master’s thesis were already sown during my bachelor’s thesis, where I had implemented word embedding for sentiment analysis. This realization was nothing short of a revelation.
Mas Aria Ghora’s suggestion to recycle my bachelor’s thesis implementation for my master’s thesis was a turning point. He proposed a simple yet profound idea: why not apply a similar approach to risk classification, specifically in the context of the capsule hotel industry? This idea not only resonated with me but also provided a clear and focused direction for my research. It was as if a fog had lifted, revealing a clear path forward.
With this newfound clarity, I began to delve deeper into the specifics of my thesis. For the dataset, I can use the feedback dataset that can be obtained from user review platforms like Playstore and AppStore. The plan was to use BERT, a pre-trained word embedding model, in combination with Random Forest, to classify risks from the dataset. This combination promised to leverage the strengths of both models, offering a robust solution for risk classification. The technical challenge was significant, but the prospect of contributing meaningful insights to the field of risk management was a powerful motivator.
In the days following our discussion, I worked tirelessly to compile these ideas into a coherent thesis proposal. The process was intense, involving countless hours of research, writing, and revision. However, the clarity of the concept and the support from my advisor and seniors made the task manageable. By the time I submitted my proposal, I felt a sense of accomplishment and anticipation for the next steps.
The culmination of this effort was the thesis proposal exam on February 20, 2023. I was both nervous and hopeful as I presented my ideas to the panel. The moment they approved my proposal was a milestone in my academic journey. It was a validation of the hard work and the collaborative effort that had gone into shaping the proposal. With the successful defense of my thesis proposal, I was finally ready to embark on the full thesis research, armed with a clear plan and a supportive network.
In retrospect, the journey from confusion to clarity was a transformative experience. It taught me the value of seeking guidance, the power of collaboration, and the importance of perseverance. As I move forward with my thesis research, I carry with me the lessons learned and the confidence gained from this journey. The road ahead is still challenging, but with a solid foundation and a clear direction, I am ready to tackle the complexities of machine learning and risk management.
After successfully defending my thesis proposal, I was allotted approximately seven months to complete the full document for my master’s thesis. While the thesis itself was undeniably challenging, the real test lay in balancing my academic responsibilities with my full-time job as a front-end engineer. The dual demands of work and study often felt overwhelming, but they also taught me invaluable lessons in time management and perseverance.
The first three months presented a significant hurdle: finding a suitable dataset for my research. My initial attempts at data scraping from the Playstore yielded reviews that were too technical and narrowly focused on the capsule hotel app itself. These reviews lacked the broader context I needed, such as service quality and overall guest experience. Just as frustration was setting in, my previous employer’s Technical Operation (Tech Ops) manager provided me with a dataset that perfectly suited my needs. This unexpected support was a crucial turning point in my research.
My previous workplace proved to be incredibly supportive throughout this journey. I was candid with my manager about the struggles I faced balancing my master’s thesis and work responsibilities. To my relief, they were not only understanding but also accommodating, allowing me the flexibility to manage both commitments effectively. Their support was a cornerstone in my ability to persevere through this challenging period.
For the technical implementation of my thesis, I chose Python as my primary programming language and utilized Google Colab to run data training processes in the cloud. This setup was essential given the resource-intensive nature of the tasks, which would have been impractical to run on my local device. Balancing my day job, which involved coding in TypeScript, with my thesis work in Python, was a daily juggling act. Transitioning between programming languages and contexts each day was mentally taxing but ultimately rewarding.
As someone relatively new to Python, I faced a steep learning curve. Understanding the mathematical and statistical foundations necessary for my model, along with the appropriate Python syntax and libraries, was daunting. Fortunately, I discovered fast.ai, which offers a machine learning crash course specifically designed for software engineers. This resource was immensely helpful, providing me with the foundational knowledge and practical skills I needed to advance my thesis work.
After seven grueling months of hard work and relentless determination, I finally completed my thesis on November 27, 2023. The culmination of this effort was the thesis defense, where I presented my work before a panel of three doctoral-level lecturers. It was an intimidating experience, but one that ultimately affirmed the value of my hard work and dedication. Successfully defending my thesis was a profound milestone, marking the end of a challenging yet incredibly rewarding journey. The experience taught me the power of perseverance, the importance of a supportive network, and the immense satisfaction that comes from overcoming formidable challenges.
The euphoria of successfully defending my thesis was short-lived. Just weeks after my thesis exam, my previous workplace announced a series of layoffs due to significant business changes. To my dismay, I was among those affected. It felt like an emotional rollercoaster—having just overcome the substantial challenge of completing my master’s thesis, I was now faced with the daunting reality of unemployment. The announcement came in December 2023, and I had to leave the company by the end of January 2024.
In the midst of this turmoil, a silver lining appeared. An old friend from my bachelor’s days reached out to me via Twitter, offering a part-time frontend role at his AI startup. We had known each other since 2016, and his offer was a lifeline. He graciously allowed me the flexibility to work part-time while I completed my master’s studies and searched for a new full-time job. This opportunity provided both financial stability and a sense of purpose during a challenging period.
Despite the job setback, my academic journey was far from over. To achieve my master’s degree, I needed to publish a research paper in addition to passing my thesis exam. From January to April 2024, I dedicated myself to finding a suitable publisher for my research. The process was fraught with rejections from prestigious publishers like IEEE and Data Brief, as well as several other international journals. Each rejection was a setback, but it also spurred me to refine my paper further.
Submitting a research paper is a meticulous process, unlike submitting a CV for job applications. Parallel submissions to multiple publishers are often considered "plagiarism," as the same draft cannot be reviewed by different publishers simultaneously. This meant I had to submit to one publisher, await their review, and only then consider other options if necessary. The waiting periods were nerve-wracking, adding to the stress of my already precarious situation.
Fortune smiled upon me once more when my lecturer, who also served as the Head of the Master of Information Systems Management Study Program, advised me to consider local publishers. Following his advice, I submitted my paper draft to Tunas Bangsa, a local publisher under Jurnal Sinta 4. To my immense relief, I received a Letter of Acceptance, and my paper was finally published on April 28, 2024. This publication marked the culmination of a long and arduous journey, validating the hard work and resilience that had carried me through multiple challenges. You can check my published paper here.
The journey from thesis completion to research paper publication was a testament to the unpredictable nature of life and the importance of perseverance. Despite facing significant setbacks, including job loss and multiple rejections, I managed to navigate these challenges with the support of friends, mentors, and my own determination. This period of my life taught me the value of adaptability, the importance of a supportive network, and the satisfaction that comes from overcoming adversity. As I look forward to new opportunities, I carry these lessons with me, ready to face whatever comes next.
So, that’s my story of pursuing a master’s degree while working full-time during a challenging period in the tech industry. The journey was filled with numerous ups and downs, many of which I couldn’t detail here because there’s simply too much to tell. However, I hope you understand the essence of my experience—the resilience, the support, and the relentless pursuit of my goals despite the odds.
As I look towards the future, I admit that uncertainty still looms. Over the past four months, I’ve had several job interviews, and a recurring question has been, "Why did you choose a machine learning topic for your thesis when you’re a front-end engineer?" My response has always been, "I wanted to confront my fears head-on. The challenge was there, and I chose to embrace it rather than remain in my comfort zone." This mindset has driven me to explore new territories and grow beyond my initial expertise.
With my academic journey now behind me, I am fully open to new opportunities without the worry of academic interruptions. While my primary focus remains on working as a web developer, particularly in front-end engineering, I am also eager to explore other areas such as Rust, machine learning, and AI. These fields intrigue me, and I believe that expanding my knowledge in these areas will not only enhance my skill set but also prepare me for the diverse challenges and opportunities that the future holds.