I’m happy to share a recent milestone: I’ve earned the AWS Certified Cloud Practitioner certification!

Over the past two years, I’ve mainly been working with the GCP platform, gaining hands-on experience in SaaS cloud operations. But while preparing for this certification, I got the chance to take a more systematic and comprehensive look at AWS cloud architecture and best practices.

Actually, this whole journey started from a fun experience at the AWS Summit this August. I joined one of the sessions where the speaker hosted a small competition. Luckily, I won a prize spot, and to my surprise, the prize was a 100% discount voucher for the AWS foundational-level exam (worth around USD 100!). That turned out to be the real “value for money” reward of the event. I didn’t get any other swag, but honestly, this voucher alone made the trip more than worthwhile. And that’s how I got the chance to take this certification exam.

It’s true, the AI services we use every day, like ChatGPT, Agents, or Vibe Coding, all rely heavily on cloud infrastructure.

And when we migrate services from on-premises to the cloud, it’s not just about moving things over. The design has to consider Security, Reliability, Performance Efficiency, Operational Excellence, and Cost Optimization. Doing this right requires both solid technical knowledge and forward-looking architectural thinking.

For me, continuous learning isn’t only about picking up new technical skills. It’s also about living up to my responsibility as a Team Leader. Every architectural decision, big or small, can significantly shape the future of the product and the growth of the team.

Moving forward, I plan to bring together my experience in both GCP and AWS to design more resilient and efficient cloud architectures, while also embracing the new opportunities that AI and the cloud will bring.

PS: AWS certifications need to be renewed every three years, so they’re not permanent. Time to keep learning!

Credly Link: Here

How to Prepare?​

The first step is to decide which language you want to take the exam in, because this will directly affect how much preparation you need.

Although AWS offers a Chinese version (with some extra time allowance, which is quite friendly for non-native speakers), I chose to take the English version. The reason is simple: translations can sometimes be inaccurate, and since I usually work in English anyway, I treated it as a small challenge.

Exam Scope

The preparation strategy is straightforward: repeatedly review each AWS service, understand its main purpose, and be able to distinguish between services that look similar. AWS doesn’t provide duplicate services, so if two seem to overlap, it’s important to clarify their differences.

Since AWS has hundreds of services, here’s a small trick: check the Official Exam Guide. If you can immediately explain every term listed there, you’re probably ready for the exam.

Study Materials​

I didn’t use Udemy or YouTube. Instead, I focused only on AWS official resources plus practice exams, and that was more than enough.

This course is available in both Traditional Chinese and English. However, I strongly recommend the English version, because I noticed the Chinese course content is shorter (about 7 hours), while the English version is nearly 13 hours. The English one feels more complete, rich, comprehensive, and up-to-date.

My suggestions:

• Go through the entire 13-hour course without skipping any detail.
• Complete every quiz with a perfect score (100%).
• If you make mistakes, review why and clarify the concepts.
• Take notes and keep revising them repeatedly.

Practice Exams​

After finishing the course, the next step is practice exams. They are incredibly helpful for understanding the exam style, question depth, and difficulty. Honestly, practice exams are a must!

Two sources I recommend:

• Examtopics (free, but limited number of questions)
• Passnexam (nearly 300 questions, highly recommended to go through all of them)

Exam Process​

The exam can be taken in two ways: online from home or at a testing center.

Personally, I recommend the testing center, because it feels more like a formal exam 😆. Another reason is that you only need to bring valid IDs (passport and national ID card). No extra setup is required, and the risk of technical issues (like poor internet) is eliminated.

I took mine at the 恆逸 Training Center near Nanjing Fuxing Station in Taipei. The staff were professional and experienced.

The process was:

• Check in and sign consent forms about exam regulations.
• Undergo a pocket check (no belongings allowed inside). Backpacks must be stored in lockers.
• If you arrive early, you can start the exam early.
• A staff member will guide you into a large room divided into many small, enclosed booths. Each booth has a computer, monitor, and chair.

The booth is quite small that basically it's just enough for one person. The chair is almost against the door, and there’s no space for a second person. The whole session is under video surveillance.

The exam lasted for 100 minutes, but I completed it in about 60 minutes and submitted my answers. Immediately after submission, the screen displayed the result, showing whether I had passed or not.

Conclusion​

The AWS Certified Cloud Practitioner certification is a great starting point for anyone who wants to step into the cloud world.

It provides a clear understanding of what cloud infrastructure is, why it matters, and how AWS services create value. The exam itself is designed at the entry level, so it doesn’t require prior coding or technical knowledge, yet it still gives a complete overview of the cloud landscape.

This makes it especially suitable for non-engineers, project managers, sales professionals, or anyone who has never written a line of code or had any cloud-related background. For those with experience in other cloud platforms such as GCP or Azure, it offers a structured way to quickly get familiar with AWS. And for absolute beginners, it serves as an accessible doorway into the world of cloud computing, opening up opportunities to grow further in this field.

As an experienced AI engineer deeply invested in Google Cloud Platform (GCP), I had the privilege of participating in the "Google Cloud Generative AI Training Program (Cloud AI Study Jam 2024)" organized by the Google Cloud Taiwan team in October 2024.

This intensive one-month program required completion of 25 courses, including hands-on practical exercises, with approximately 48 hours of dedicated learning time. Through the Google Cloud Skills Boost platform, I gained direct hands-on experience implementing LLM technology applications such as Google Gemini and Vertex AI Studio within the GCP environment. Beyond earning a Certificate upon completion, participants who passed Google's review process were rewarded with an attractive set of complimentary gifts!

✔️ The beginner level provided a solid foundation in generative AI, focusing on LLM principles, the training and adaptation of foundation models, and challenges such as model bias and hallucinations. Through hands-on experience with prompt design in Vertex AI Studio, I developed skills in zero-shot, one-shot, and few-shot prompting techniques, significantly enhancing my ability to optimize AI solutions for tasks like Q&A, text summarization, and information extraction in the environments.

✔️ The intermediate level explored Gemini's practical applications, where I gained valuable insights into its integration across development ecosystems. The SDLC training demonstrated Gemini's capabilities in improving efficiency throughout the tech stack, from cloud architecture, data science, network engineering to security engineering and DevOps workflows.

✔️ The advanced level challenged us with in-depth technical concepts, including image generation principles, attention mechanisms, encoder-decoder architectures, transformer models, and BERT model implementations. Alongside developing expertise in these technical components, we focused on implementing robust MLOps practices while addressing crucial aspects of AI fairness, interpretability, and privacy protection.

This program offered invaluable insights into generative AI applications, particularly with Gemini models, prompt engineering and Vertex AI on the Google Cloud Platform.

After successfully completing three distinct learning paths in the development journey and getting certified in November 2024, two months of patient waiting paid off when the Google Taiwan event team delivered an exciting package this week, which included Google Cloud stickers, a high-quality stainless steel tumbler, and a specially designed laptop backpack.

I'm grateful to the Google Cloud Taiwan team for this comprehensively designed learning opportunity. A special shout-out to Google for their dedication to developer education and innovation! Looking forward to future opportunities to advance my expertise in cloud and AI technologies.

As a developer, it was a great honor to attend the NVIDIA Developer Meet-up held in Taiwan today. The Meet-up was packed with valuable content, featuring renowned professors from academia and entrepreneurs from startups who shared their precious experiences and knowledge.

One of the most exciting moments for me was finally meeting Professor Hung-Yi Lee from National Taiwan University and listening to his live lecture. Several years ago, while pursuing my Master's degree in the UK, I watched his YouTube channel sharing content on machine learning and deep learning. Today's live lecture showcased his consistently engaging and interesting style, explaining extremely complex and profound AI principles in the most easy-to-understand manner, which was greatly beneficial.

Besides that, I also had the honor of listening to a talk by Professor Tzong-Han Tsai from Academia Sinica. Right from the start, he shared a music song created using LLM, with very interesting lyrics and melody, instantly shaking vibe up.

Of course, I also learned from other speakers who generously shared the architectures they designed using NVIDIA software, along with some interesting business ideas and perspectives.

Furthermore, I was delighted to see NVIDIA CEO Jensen Huang again, this time standing on stage, sharing some advice to the developers present. I could feel his sincere hope that Taiwan does not only focus on hardware but also keep pace with AI in the software field this time. He specifically emphasized "No Excuse!" several times. 😂

Finally, I am truly thankful to NVIDIA for hosting this AI Summit event for developers, willing to spend so much money and manpower to organize these kinds of events, which is highly admirable.

I also want to thank Wei Jing (William) Wong for his outstanding assistance in hosting this event and his perfect control of the meetup flow. Additionally, I am grateful to Solomon AI and 3D Vision for giving me the opportunity to participate in this event.

Today, I had the great fortune of attending NVIDIA CEO Jensen Huang's keynote event in Taiwan.

It was an incredible experience to see firsthand how a tech giant organizes a large-scale event.

As an attendee, I'd like to share some impressions beyond the online livestream content. From managing the queue of thousands, distributing umbrella covers on a rainy day, and providing an NVIDIA-branded shirt on every seat, every detail was meticulously handled.

Even after the event, staff collected the umbrella covers, showcasing their attention to detail.

The keynote felt like a concert with top-notch lighting effects, and the attendees were all exceptionally talented individuals. I noticed that Jensen Huang gave his speech without using any teleprompters, and his performance was truly outstanding and inspiring.

Besides, it was an honor for Solomon to be mentioned again as an NVIDIA partner. We will continue to push the boundaries and advance on the path of AI!

Next week, we will also be at the ComputeX exhibition. We sincerely invite you to come and visit our booth.

This was my second time participating in Taipei Automation Exhibition.

We showcased the second-generation APP of Solomon's wearable AI platform META-AIVI, which has fog inference computing capabilities.

In the past half year, we dedicated to developing a new and completely different workflow for users to provide a more reliable and flexible interface.

Let me highlight one of the strengths for the AI part of the 2nd generation META-AIVI. Our new AI technique (i.e., unknown counting) can rapidly count a number of objects by few labels WITHOUT training.

In particular, on the first day of the exhibition, Taiwan president Tsai visited our booth and we introduced our new product META-AIVI to her, which is our honor.

As the core developer of this product, I can say that the whole process was pretty tough and stressful, but at the end of the day, we managed to overcome all the difficulties. 😃

This was my second time participating in the Kaggle competition, TensorFlow - Help Protect the Great Barrier Reef, and the deadline was only 1 month to go. The topic domain of the competition was about object detection of computer vision.

As you can imagine, most people and teams were using the architecture of the YOLO series to train the models, and most of the discussions were about YOLO methods for optimizations.

Honestly, I think that this time the dataset was not distributed really well indeed. Many teams were ranked vary significantly between public LB and private LB.

Besides, in my opinion, I really thought that it is highly difficult to get a good result (or score/rank) if you do not have enough (awesome) hardware resources.

In short, I appreciate my teammates who have made a lot of contributions as well. I am so glad that we can win the silver medal from over 2000 teams on the public LB of this competition. 😃

I recently participated in a Kaggle competition with my friends and this was an awesome experience for me because it was my first time joining a Kaggle event.

We used our personal time to discuss strategies and investigate many methods. I learned a lot during this competition such as the DDP technique of Pytorch, processing a pretty large dataset (over 500 GB), different models with the distinguish training methods, etc.

I really appreciate my teammates who have made a lot of contributions as well. Without them, it would be impossible to win the bronze medal at Google Landmark Recognition 2021.

For me, this is a good result for those who are participating in the competition for the first time. I believe we will achieve better results in the next competition! 😃