Hustle Hub #32

New to Hustle Hub? Make sure to subscribe for more!

Hey friends,

I’m back to Malaysia recently and have a short vacation with my family at Genting Highlands. After a short burnout last year, I often remind myself of the importance of family. Work, on the other hand, is just work.

What matters most is still the precious time spent with family, not forgetting to have fun and enjoy each other’s presence. 💜

Do you often spend quality time with your family? Let me know =)

In today's issue, I’d like to share with you how to achieve product-market fit, how to build your confidence by doing, a very interesting article to build LLP applications in production, and a cool tool to build AI models quickly.

Let’s get to it! 🚀 

🛖 What's in the hub today?

• Tip: How to achieve product-market fit

• Mistake: I was not confident

• Learning: Build your confidence by doing

• Article: Building LLM applications for production

• Tool: Predibase

⭐️ 1 Tip

How To Achieve Product-Market Fit

Recently I met a founder on Lunchclub and we talked about building a product to achieve product-market fit (PMF). He is building a product with more than 150,000 active users within a few months and he shared with me 2 approaches that can be used to achieve PMF.

I found these 2 approaches very insightful and can be used literally by anyone, especially if you’re building a product (or MVP) to test out your startup idea.

Approach 1: Market → Product

With this approach, you first:

• Identify the problem and the assumptions you have

• After that, you conduct user interviews to validate or invalidate your assumptions to see if you’re solving a real problem.

• Finally, once you’ve validated your assumptions and the problem statement, you start building the product (solution).

• This approach is commonly used if you’re not personally facing the problem, but rather you want to help others solve their own problems.

• Therefore, you start by finding the market, then build a product.

This is the approach that we used at Staq.

• When we started building Staq, we learned about the problem from customers (market), then we built the solution (product).

• This is useful since we don’t personally encounter the problem, hence we shouldn’t build the product first, then go find the market.

Approach 2: Product → Market

With this approach, you first:

• Build a product (i.e. MVP), launch it, test the market, and iterate.

• Use this approach if you’re personally facing the problem, and you already know what the product (solution) looks like to solve the problem.

• But this approach is a bit risky because you have not validated if your customers face the same problem you’re facing. Hence, quickly building and launching MVP help to reduce the risk without incurring high costs.

This is the approach used by my friend who is building his startup’s product:

• Since they personally face the problem of highlighting notes on websites for learnings, they didn’t start with finding the market.

• They first built the MVP, launched it, and then only started identifying their ideal users (researchers) and market.

• This approach worked for them because they understand deeply about the problem they want to solve, so they can just go straight into building the product (solution).

⚠️ 1 Mistake

When I first started out, I had low self-esteem. I didn’t think I was good enough. I had low self-confidence.

Because of that, very often I didn’t speak up because I was afraid of saying the wrong things or how people would think of me if I said something wrong.

My low self-confidence led to suboptimal communication with others. And suboptimal communication led to misunderstanding at work.

Not good.

🧠 1 Learning

After noticing my weakness, I decided to fix it.

I started by challenging myself to do something that I’ve never done before, and to actually achieve it. Because I believe that confidence comes from repeatedly doing and achieving what you wanted to do.

For example, before we pivoted to Staq, we were actually building a mobile app to help food delivery riders consolidate their income across various platforms (Deliveroo, Foodpanda etc.) — you get the idea.

We had limited resources to hire developers to build the mobile app for us, so I self-learned Flutter from Udemy and built a mobile app as our MVP within 2 months — and we did it!

Although we shut this idea down soon after our launch, but learning and building a mobile app from scratch boosted my self-confidence tremendously to do anything in the future.

🧠 Here’s what I’ve learned:

• Self-confidence is not born, but built.

• Try to challenge yourself to do something that you’ve never done before (outside your comfort zone), and to actually achieve it. By doing so, you self-confidence will be improved significantly — trust me.

• Always have a growth mindset to continuously learning. Do not be afraid of other people’s judgement. The more you don’t care about what people think about you, the more confident you’ll become.

📚️ 1 Article

Building LLM applications for production

Ever since the launch of ChatGPT, many developers want to build LLM applications, but not many of them know how to build LLM applications for production.

That’s because building LLM application for production is very different from building for development.

When Chip Huyen wrote an article about this topic, it just caught my attention. Needless to say, I learned a lot from her sharing and just thought of sharing with you too — especially if you’re also looking to build an LLM application.

Check it out and you’ll know what I mean.

Have you read this article? What's your thought on it?

🧰 1 Tool

Predibase is a low-code AI platform where you can easily build custom ML pipelines declaratively or instantly query popular LLMs hosted within your own environment.

Finetuning models is just as easy, and deployment is only one click away. Start putting AI projects into production in minutes, not months.

Building an AI model in a Jupyter notebook is easy. However, when it comes to building and deploying an AI model is production, things can get complicated easily due to complex MLOps, models monitoring, features improvement etc.

This is where Predibase can help you to build and deploy AI models in minutes, not months.

Have you tried it out? Would love to know your thought on it!

🔥 Hustler Spotlight 🔥

👋🏻 How would you introduce yourself?

Hi everyone! I am Cornellius Yudha Wijaya, a data scientist and avid data writer from Indonesia who loves to learn about various things. Despite initially graduating with a degree in Biology and being a biological researcher, I have accumulated over 5 years of experience in the data science field. I enjoy sharing my knowledge and experiences in transitioning into the data field with everyone.

👀 What’s your day to day like in your current role as a Data Scientist Assistant Manager at Allianz?

I would say 50% of my role is about communicating with relevant stakeholders and collaborative teams. Data initiatives are primarily about solving business problems, so I dedicate much of my time to speaking to the business. This involves defining their needs, identifying the right data sources, confirming Key Performance Indicators (KPIs), and more. If the project is ongoing, it's about keeping up with the business, communicating results, and determining the next steps.

About 30% of my work is technical, including querying data, cleaning data, modeling, and maintaining the model. This also includes documenting the project and sharing the results with the team. I also maintain MLOps within the company, so finding the right solutions is part of my daily activity.

Of course, I am not working alone. The support of my team is indispensable, and being able to communicate effectively with them is critical.

Lastly, 20% of my time is reserved for ad hoc requests and personal development. Occasionally, requests arise outside of the intended initiative, helping out the team, or any other tasks. Furthermore, ongoing learning is essential for career growth, so I also set aside time for that.

⭐️ What has been the biggest highlight of your career so far?

I take pride in all my achievements, from seemingly small feats such as completing my thesis to making a career switch into the data field.

However, if I chose the most significant, it would be implementing a Cloud-based data analytics environment within my company. Prior to this, my company relied on an on-premise setup and depended heavily on additional servers to scale up data needs, with no clear automation process in place. I initiated and spearheaded the shift to cloud-based analytics a year into my role. I am proud of this achievement, but what stands out most to me is my perseverance in not giving up on the idea and effectively communicating it in a way that engages the stakeholders.

Outside of my professional life, a highlight for me is that my articles have been read over two million times, and I've connected with thousands of people on social media. It still feels surreal when I think back to just five years ago when I was unsure about my career path.

🚀 What's a data or AI trend you're watching this year?

‍Large Language Mode, of course. It’s a trend that has changed the world; nobody should miss that.

I'm also following the MLOps trend this year, as I believe automation will become increasingly important.

💼 What advice would you give someone starting out in Data Science?

Never stop learning. Data science is a vast field that's growing each day. What you learned yesterday might not be relevant tomorrow. Everything is rapidly evolving in the data science field. That's why it's essential to understand the foundations, as they will benefit your career in the long run if any new concept is coming.

However, you also need to understand your preferences and limitations. While learning all aspects of data science is beneficial, having one or two favorite concepts, algorithms, or methods could motivate you during your work. Also, recognizing your limitations is crucial for managing your mental expectations.

Soft skills are as important as technical skills, especially communication and presentation. You might be able to create the most advanced machine learning model, but if you can't communicate and persuade stakeholders, it's all for nothing.

Lastly, always have fun with the projects you undertake! It might seem counter-intuitive, but enjoy what you do because you deserve it.

🤯 What’s the most important career lesson you wish you’d learned earlier?

One might think that their future is set after finishing a university degree. However, that's not necessarily true, and you can always become anything you want if you dedicate your time to learning.

Networking is essential. My future would have been different if I hadn't networked enough. I got my job because I networked. I obtained my freelance gig through networking. I'm able to write here because of networking. If I could turn back time, I would start networking as early as possible.

Don't rush things. Learn at your own pace because rushing only leads to messes.

🧠 How would you learn Data Science if you had to start over?

Start with the fundamentals. I recall diving into data science and immediately jumping to advanced coding and machine learning, which only confused me. I wasted too much time because I had to return to learn the fundamentals. So, don't waste your time. The term "fundamentals" here refers to the basics of data science: math, statistics, and programming. These foundational elements will assist you in your data science career. Get curious all the time.

Having friends or a network with whom you can learn is beneficial. Sometimes, there might be problems we can't solve alone or concepts we don't understand. Having someone to learn with can make understanding these concepts easier.

Document everything. Maintaining neat and thorough learning materials can be extremely helpful, no matter what stage you're at in your career. If necessary, post them on the blog or the web.

🔥 Where can we find your amazing work?

You can always find me on my LinkedIn or sometimes my Twitter.

I also actively write in my Newsletter: Non-Brand Data, sharing data science latest news, tips, and daily life update.

Occasionally, you can find me writing on my Medium blog as well.

📚 What's your favorite book?

• Originals: How Non-Conformists Move the World

• Data Science for Business: What You Need to Know about Data Mining and Data-Analytic Thinking

• The Selfish Gene 

• Data Science on AWS: Implementing End-to-End, Continuous AI and Machine Learning Pipelines

• Practical Statistics for Data Scientists: 50+ Essential Concepts Using R and Python

🚀 Whenever you’re ready, there are 4 ways I can help you:

1. Book a coaching call with me if you need help in the following:

2. Promote your brand to 1,000+ subscribers in the data/tech space by sponsoring this newsletter.

3. Watch my YouTube videos where I talk about data science tips, programming, and my tech life (P.S. Don’t forget to like and subscribe 💜).

4. Follow me on LinkedIn and Twitter for more data science career insights, my mistakes and lessons learned from building a startup.

That's all for today

Thanks for reading. I hope you enjoyed today's issue. More than that, I hope it has helped you in some ways and brought you some peace of mind.

You can always write to me by simply replying to this newsletter and we can chat.

See you again next week.

- Admond