• I found meaning in work.
• I found love again.
• I found good health.
• I found good food spots.
• I found the right boundaries for my relationship with my parents.
• I found my love for listening to music again.
• I found a deeper love for reading.
• I found a passion for writing online via X[dot]com.

But I also lost a lot of things:

• I lost my love for playing guitar.
• I lost my youthful energy.
• I lost my friend groups.
• I lost my love for lifting.
• I lost my love for sports.

Of the things lost, I hope to pick these back up:

• Lifting
• Guitar
• Running

Reflection of deeper things

Friendships

The meaning of friendships changed for me ever since the loss of my core group after the breakup with my ex from college in Oct of 2024. I decided to guard my heart from ever getting as attached as I was to those friends. Today is January 1st, 2026 and my guard is still up, but I have found some new friends as well as revisited some old ones. I'm thankful for everyone currently in my life but never looking for new relationships openly. I hope that I can carry this disciplined mindset of friendship into 2026 as well.

Career

A year ago I was crashing out about my position at Amazon due to my feeling of being stuck in a role that wasn't letting me grow in the ways that I wanted to. Today I'm at an early stage startup with coworkers that I can easily spend 10+ hours with at the office while losing track of the time. Ironically, I feel way more stable at this startup position than I did at my big tech job with Amazon. I think it's because I finally got over the imposter syndrome I had from never being part of a full time gig outside of big tech. I used to have a fear that I would never be able to code anywhere else except in big tech, but now I feel like I could truly bring my talents to any team and just build cool shit, the way that I always wanted my career to look like. I also feel a tremendous amount of lift in my craft of writing high quality traditional software rather than contributing to an unmaintainable codebase as I did in my MLE role prior.

Love

Not only did I get over my breakup last year, I also found love again with my current gf who has shown me that I am capable of loving and being loved, something that I thought I'd never be able to experience again just a year ago. I reflected a lot in 2025 on how I could take the things from my last relationship and become a better version of myself, but I realized during the tail end of the year that it led to a bad habit of self loathing. One thing that I used to regret a lot from my last relationship was how much I hurt my ex by never being in the present moment and always trying to optimize my life in order to level up. My overcorrection of this manifested into constant self evaluation and self loathing. My current gf often has to snap me out of the state that I slip into when I am being far too critical of myself. Unfortunately, the desire to improve myself from my last relationship has bled into a circular mindset of finding flaws in myself in all areas of life, from career to love life, and fixating on them until I correct them, often removing me from the present moment and trapping myself in an echo chamber of negative thoughts. While I do believe that self improvement is a great thing that many people could benefit from, there is clearly a deep end that should not be crossed that I find myself crossing a lot. In 2026 I hope to find peace with who I am and self correct in a more loving way.

Music

I played a lot of guitar from 2021 to 2024, even posting over a hundred TikToks of myself playing. Last year I stopped posting on social media completely and in the last few months of 2025 I even stopped my daily habit of playing everyday. The reason is layered, from feeling like I will never be as good as my heros to feeling like I could spend my time leveling up in areas of life that will lead to financial gains (primarily my software career). In 2026 I plan on playing more guitar at lower stakes, primarily focusing on playing for playing's sake rather than for performance.

Fashion

Another big form of expressing myself has always been fashion. Nothing feels better to me than dressing up in a fire fit and walking through the streets of NYC or on a date with my lovely gf. Having it on makes me feel like I'm wearing my persona, allowing people understand me without really knowing me. That's a really important thing for someone like me that isn't really outspoken. I used to thrift a lot and did not care if I spent a few hundred bucks on my wardrobe every month. In 2026 I want to thrift a lot more again rather than buying expensive pieces (no new chrome hearts pieces in 2026 hopefully).

Conclusion

Life is very turbulent and 2025 has proven to be nothing short of a flight through heavy storms. I am walking out with fresh perspectives, new friends, and new beginnings.

Life can't be all about providing value

Performance reviews are not a sustainable metric for life outside of work. During Thanksgiving weekend, I found myself feeling meaningless because I wasn't working and providing value at my startup. It prompted me to pry into the other areas of my life like my relationship, my family, and my hobbies. Upon doing so, I realized that I only really feel happy when I am performing well in worldly standards across all of these different areas of my life.

Am I being a good boyfriend and providing?

Am I being a good son to my hard working parents?

Am I playing guitar better than most people?

These are important questions to ask once in a while, perhaps during yearly reviews, but I find myself asking these questions everyday. As a result, on days where I'm supposed to be resting I feel down and end up feeling more tired than a monday morning.

Be a high achiever. Be great. But do not try to perform your way through life. That's a sure fire way to burnout and unhappiness.

The solution?

I can't say that I've found the perfect solution to this existential problem. What I do know is that life isn't some circus act where I have to constantly balance myself on a tight rope in front of my peers, family, and friends. The best thing I can do right now is remind myself that I have more to my identity than just performing well, and hopefully you can do the same for yourself as the holiday season continues to put a dent on our productivity.

The four different types of MLEs (from my experience)

1. Data focused MLE

If you're working on an ML service, you're likely dealing with big data. While you might not be responsible for populating tables (data teams usually handle that), you definitely know how to navigate data sources and understand how the data aligns with business requirements.
Almost every MLE needs to know how to ingest, clean, and prepare data for ML services.

The data focused MLE works on:

• building and maintaining data pipelines
• ensuring data freshness and quality in their ML pipeline
• writing ETL jobs (think spark, airflow, step functions, etc.)

2. Infra focused MLE

Without the proper infrastructure, your ML system will not survive production traffic unless you are working on a toy project with no users.

Depending on your business requirements, you will need to make key decisions about the infrastructure for your ML model that directly affects how your customers experience your application and how your model performs predictions.

For advertising for example, latency is very important as ads should be available in near real time and any delays in ad data freshness will lead to "stale" impressions where the relevance of the ads are low for users. In this case, you need to build a real time inference service that holds your trained ML model that is long running. This leads to you deciding to use a long running ECS with a large instance size so that you can load things on memory in order to perform inference very fast with any artifacts being readily available on disk.

For a streaming TV service you might only need to have fresh recommendations based on user activity every few days or even weeks. This calls for a batch inference service which calls for reliable offline jobs that run on some recurring schedule but gets rid of the need for a long running service.

This is the focus that I've been in for a while and I truly love it. It requires strong distributed compute knowledge and although you don't technically need to know the science specifics of the actual models, the strongest infra MLEs have strong foundational ML knowledge.

3. Model focused MLE (applied scientist)

In the big tech world you will often call MLEs that work on core modeling "applied scientists" and these are the MLEs that most people think about when they think of people working in ML.

In reality, applied scientists are yet just another subgroup of people working in ML that also have a strong grounding on software engineering skills with applied ML knowledge.

They may work on novel papers for their teams, but for the most part the true goal of an applied scientist is to create revenue with their ML models.

I've seen most model focused MLEs not really work on "productionizing" ML services and instead focus on doing offline analysis on their experiments in order to maximize revenue or other important metrics (like CTR, relevance, etc.).

These MLEs are often VERY up to date on the latest research being done in the topic of their team's specialty so if you're on a robotics team then expect to be really up to date with DRL while if you're on an advertising team be ready to be at the edge of NLP and LLM research.

The applied scientist might work on:

• offline analysis in notebooks (sagemaker, jupyter, etc.)
• prototyping quickly and often without always writing production grade code
• sometimes contributing to papers for the company
• driving revenue gains with their carefully constructed models
• frequent experiments and sharing results with shareholders

4. Full-stack MLE

Finally, this is a new term I am hoping to coin for those MLEs that seem to do everything from the data cleaning to the modeling to the scaling.

I interviewed with a startup a while back and the role required me to work on everything as I would've been the only MLE on the team.

This seems to be more common in the startup world and in general I would say that the full stack MLE prioritizes making things work.

Making things work is a very software engineer / indie hacker - centric concept but even for the MLE there are times when you just really need to make shit work for your team.

I would say that if the other categories of MLEs go super deep, the full-stack MLE has the widest breadth of knowledge across the full ML stack.

Some readings for you:

I will just leave you with one thing for the weekend that might pique your interest:

• ​https://fullstackdeeplearning.com/​

The resource above gives a good overview of how broad the end to end ML workflow is.

If you like content like this, check out my MLE newsletter which focuses on real experiences from being a MLE: list.raymondyoo.com

STOP. You probably have 15 different ML textbooks on your laptop and 20 bookmarked courses.

Before you start one of them again, you should think about what you really need. In fact, even before that just think about what you really want from your career.

If you want to be a theoretical NLP researcher then it might make sense to finish that linguistics textbook but if you just want to be "smart enough" to work in ML then for the love of God PLEASE do not start another theoretical course.

Theory vs application

ML is not all theory. It's easy to think that you can just read all the textbooks available on DL and go through every single course out there but ask anyone actually working in ML and they will quickly tell you that it is unnecessary to do read more than 1 or 2 theoretical textbooks / go through more than 1 or 2 theoretical courses.

ML by itself is just a sub topic within CS, similar to automata theory, tree algorithms, and cryptography. When applied though, it's a powerful tool that brings in billions of dollars of revenue to the biggest companies and startups alike.

The scientists working on the models that power these ML applications have more often than not gone through many years of schooling as they honed their craft in doing research, but there are also a large amount of people that were just regular software engineers that ended up working in the ML space such as yours truly.

The speed run guide to ML =/= more courses

If your goal really is to just have all the theoretical knowledge in the world about ML then sure go read all those 10 textbooks you have saved on your laptop (although there will be lots of overlap). However, if your goal is to become someone who works on ML services then just finish 1 textbook or a single course on the foundations of DL and then start applying that knowledge on top of the software engineering skills you already have.

MLEs just apply ML to software applications. That's it. No black magic and no crazy scientist shit.

I wasted sooo much time reading and rereading textbooks on the same topics over and over again but at the end of the day, my true acceleration in my ML career came from my applied work of building and scaling large ML systems that often times used the most basic classical ML models such as random forest models.

The theory knowledge definitely helps in conversations with science teams and understanding business requirements, but the rest of the work comes from applied knowledge.

Anon, you are smarter than you think. Don't pick up another theoretical course and instead if you are going to go through some learning resource, look for applied courses/resources.

Here are some cool applied resources I can vouch for:

• ​the ML systems book​
• ​the ML databrick book​
• ​Full stack DL​

If you want to be a good MLE then make sure to check out my email list cuz they heard it all first: https://list.raymondyoo.com

Having those publications under my belt was helpful in behavioral interviews and gave me confidence in my foundational DL knowledge, but if I’m being honest, it didn’t directly help me land a career in ML. In fact, even after publishing those two papers, I realized that if I wanted to continue doing research, I would need to pursue a PhD program, then a postdoc, and eventually either join industry or become a professor.

When I really thought about it, the idea of barely getting by financially as I did research—hoping that maybe I’d get a good job after all of that—just wasn’t worth it. Luckily, I found an easier way to start a career in ML.

Becoming a SWE

When my research side quest ended, I joined my peers in pursuing software engineering internships, and luckily, I landed a great role at a big tech company. After graduation, I started working there full-time.

I wasn’t initially working on ML. In fact, I was doing pretty standard full-stack web development—work that any bootcamp graduate could handle with sharp React skills. It was fun to build things, but my desire to work on ML was still there, so I began looking for ways to make it happen.

With a working knowledge of theoretical ML still fresh in my mind, I constantly asked for opportunities to work on ML projects. Fortunately, an ML project finally became a high priority and although it a small-scale project, it taught me that ML is not just about theory. In fact, almost none of my theoretical knowledge from research helped me. Instead, I relied heavily on my foundational software engineering skills: making sound infrastructure decisions, designing scalable data pipeline jobs, and ensuring availability and reliability in our ML service.

MLEs are just SWEs that do some extra things

After working on that ML project, I realized that MLEs are basically SWEs with domain expertise in the various components of ML services—data engineering, modeling, infrastructure, and so on.

Using this new understanding and the project under my belt, I was able to transfer teams internally, becoming a full-fledged MLE at my company. Now, it’s my passion to help others break into ML without being misled into thinking that research is the only path.

My name is Ray, and I want to help as many people as possible work in ML in the easiest way possible and share my insights as I navigate my career. If you like content like this, check out my email newsletter too cuz they heard it first!

list.raymondyoo.com

Wait, what?
Before you "actual creatives" roll your eyes and start crafting your takedown tweets, just give me a second. There's this crazy misconception floating around that programming is just about slapping keys like a caffeinated keyboard warrior, fueled by Cheetos and a dangerously low shower count.

The Real Deal: Code Is Just the Beginning

Let's get one thing straight: programming is so much more than smashing out lines of code. Sure, there are always going to be those folks in any field who fit the lazy, unhygienic stereotype, but the idea that this defines software engineers? Total myth. The reality? A true developer spends about 80% of their time in the planning phase—sketching out ideas, drawing up mockups, researching, talking to users, and experimenting. The actual coding? Maybe 20% tops.

If you're a fresh-faced computer science student, still traumatized by those 3,000-line programs that barely passed the test cases, I get it. But those late-night coding marathons? They're just a warm-up, a primer in logical thinking. Once you step into the real world, whether you're building something for fun or cash, you'll realize that coding is just one tool in a much bigger kit. It's the creativity, the ability to envision something unique, that turns lines of code into something truly amazing.

Creative Coding: Enter the Indie Hackers

Here's where it gets interesting: ever heard of an indie hacker? If not, you're missing out. An indie hacker is like a modern-day renaissance coder—self-sufficient, not chasing VC dollars, but instead launching their software products solo, relying on their own skills and creativity.

The secret sauce that sets successful indie hackers apart? They’re not just grinding code 24/7. They’re reading, listening to podcasts, dabbling in design, photography, music—hell, even building personal brands with blogs and vlogs like this. :) The sweaty, Cheeto-fingered stereotype? Nowhere to be found. The real deal is a coder who steps away from the keyboard, learns something new, and brings that fresh perspective back to their work. That's how you build something people care about.

The Takeaway: Be More Than Just a Coder

So, if you're a programmer, here's my challenge: step outside. Travel. Read. Pick up that guitar you've been eyeing. Live a little and let the world outside of code inspire your next project. Don’t shy away from calling yourself a creative—own it. Your code can be art, but only if you let your creativity off the leash.

And if you're not a programmer? The same rules apply. Whatever your profession, push the boundaries, think outside the box, and find those creative solutions that make your work stand out. If coders can be creatives, so can you.

Still skeptical? Check out the Indie Hackers Podcast and see for yourself. SWEs might just be some of the most creative people out there, right up there with artists, musicians, and anyone else who dares to think differently.