Bridging the Gaps: My 2026 ML/AI Engineering Roadmap
The Imposter Syndrome is Real
My brain is constantly telling me that I need to understand all of the underlying concepts before I can confidently claim a title. Currently, I describe myself as an ML/AI Engineer, and honestly, it sometimes feels odd to say it out loud. What makes it even more real are job interviews that sometimes feel more like interrogations about specific terms.
Don’t get me wrong—I am not a beginner in this field. I have solid experience under my belt, but I know I have gaps in my knowledge. By the end of 2026, my goal is to completely eliminate that doubt. I am prepared to do whatever it takes to become an expert-level ML/AI Engineer, and this post is all about how I plan to get there.
Finding the Right “Why”
Like many people, I struggle to find the motivation to learn new things. I can’t just read a textbook; I need a purpose and tangible outcomes. To keep myself motivated, I realized my learning needs to be driven by one of three things: building a project I am interested in, preparing for a specific certification, or a genuine interest in exploring new knowledge.
Here is how I broke down the pros and cons of each approach:
| Approach | The Pros | The Cons |
|---|---|---|
| Building a Project |
|
|
| Certifications |
|
|
| Custom Upskilling |
|
|
Because I want to maximize my impact and hold myself accountable through this journey, I set up this personal blog to document my progress!
Where I Stand Today
To get to where I want to be, I first needed to honestly assess what I already know and what I need to learn.
What I am confident in:
Software development in Python
Standard Machine Learning concepts
Containerization & deploying containerized apps to the cloud
Local model development and deployment
Generative AI applications
API/Service Integration (SOAP, REST, gRPC, WebSockets)
CI/CD pipelines
The Gaps I need to fill:
Cloud architecture (the big picture)
Data engineering (specifically PySpark and Databricks)
MLOps (orchestration, feature stores, end-to-end lifecycle)
LLM expertise (different architectures, K-V cache, fine-tuning, quantization)
The 2026 Execution Roadmap
With my gaps identified and my goals set, here is the concrete roadmap I am following this year to level up from competent to expert:
Master Cloud Architecture (AWS SAA-C03): AWS is the oldest and most widely used cloud provider. The Solutions Architect Associate (SAA-C03) certification is the perfect fit for my needs right now—it provides deep, foundational cloud architecture knowledge without being absolute overkill like a professional-level exam.
End-to-End Custom ML Deployment Project: I will build a custom ML deployment showcase project from start to finish using open-source tools. I’ll follow this up with an in-depth article on this blog detailing exactly how I would deploy the architecture to the cloud.
Learn the Fundamentals of Data Engineering: I will be diving into PySpark and Databricks to shore up my data engineering skills, ensuring I can handle the data pipelines that feed my models. This goal is not highly specific yet, but I feel that deep expertise here isn’t strictly necessary right now. I might come up with a small project later if needed.
Achieve the GCP Machine Learning Professional Certification: To round out my cloud and ML knowledge, I will tackle the Google Cloud ML Professional exam.
Hands-on LLM Fine-Tuning and Quantization: Finally, I would like to learn how to fine-tune and quantize Large Language Models locally.
This year is all about bridging the gap between knowing how to build things and truly understanding the systems that power them. Follow along as I check these off the list!
Even though I am not sure that I will accomplish it all, I think this roadmap will steer me in the exact right direction.