In this “post” I just wanna ask how to get into embedded systems for those who’ve worked in it as a hobby or as a professional. Maybe there are specific projects you have in mind, or groups I can join.
I’ve been professionally building and architecting software and software systems (in the web development space) for about 6 years now, but I want to delve into more advanced, lower-level concepts in an entirely new engineering landscape. I’ve delivered solutions in many high-level programming languages, and learning a lower-level language like C or C++ won’t be an issue.
Does anyone have any thoughts?

Micah Norwood

A web developer with 6 years of professional experience seeks guidance on transitioning into embedded systems development. They're looking for project recommendations, community groups, and learning paths to explore lower-level programming concepts, with willingness to learn C or C++.

How to Get Into Embedded Systems

I think a good answer requires a bit more information about the field you desire to get into. For me, “embedded systems” can mean anything from Robotics, Industrial Automation (PLCs, IO boards), smaller more specialized IoT devices or system integration up to PCB design and FPGA programming.
I can provide some insight into firmware for industrial automation. More specifically PLCs. These are essentially rugged little computers capable speaking a bazillion different protocols and offering safety guarantees in building-, manufacturing, marine- and energy-automation. C is fundamental for firmware but sometimes even that is too bloated. For me, knowing Assembly language is important as well, as sometimes software abstractions like kernels are toooo slow. It is also necessary to understand disassembled debug builds and learn the processor architecture. Tooling in general tends to be rather old. That makes it more difficult to work with than it sometimes needs to be. In terms of languages I also need to know specialized languages like Structured Text or FBD (used to program PLCs by system integrators). While not low level it poses significant limitations for the programmer and has some rather unusual concepts (like Function Blocks which are essentially callable structs). In robotics you might encounter even weirder languages like KRL, used to control kinematics. AI does not tend to be very useful in this field, specialized langues like IEC 61131-3 are hardly present in training data and from my experience AI is very cumbersome to integrate properly into 30 year old codebases with specialized requirements for safety, formatting and testing.
I got into the field as a working student applying for an IT job at a large company. By chance I landed in the PLC firmware team. Takes while to get into as it requires a lot of knowledge about networking (Ethernet/IP, Fieldbusses and Serial), operating systems, hardware (CPUs, microcontroller and memory) as well as cryptography (when OpenSSL is too heavy for than one tiny core). It can be a very wide field and often different solutions are applied as opposed to the world of the cloud. Software is often decades old and it shows. Rust for example, while promising, might take another decade to truly be established by big players in industrial automation. It’s a rather slow moving field where reliability and correctness matters, which doesn’t mean there are no janked, half baked solutions around.
I can’t tell you exactly what will land you a job or opportunity but I will tell you about some stuff I think could help in job interviews or provide ideas to explore as a hobby:
<ul>
<li>
Small projects with hardware (Arduino, Raspberry PI). Build an LED Matrix and display a GIF. Automate something small at home. I know a colleague at work who’s hooking into the CAN network of his scooter to get diagnostics out of it (this is too advanced for me though).
</li>
<li>
Be interested in processor architectures and dig into hardware abstractions like memory virtualization. Maybe write a “Hello, World!” in assembly to see whether embedded systems are really as interesting as you thought (I certainly do).
</li>
<li>
Networking is really important. Automation is not going slow down in the next years and even the smallest IoT devices can speak multiple protocols. Some even offer real time performance, many new protocols (everyone since 2000) are based on Ethernet (PROFInet, EtherCAT, OPC UA, …).
</li>
<li>
A background in Web Development/System design can help. Visualization Panels sometimes are nothing more than Webpages loaded from a PLC nearby. PLCs often transmit data to the cloud for predictive maintenance or cost optimization and automation systems these days can have complex networking topologies.
</li>
<li>
Build/Configure your own hardware. Building your own PC from parts and learning about physical interfaces is a decent start, albeit costly.
</li>
</ul>
I know, this text is long and may not give you anything you looked for. If you have any questions feel free to ask. I would be interested to hear if you are interested in a certain field like Robotics, Industrial Automation, IoT or system integration?
I think this would help people a lot with providing useful information rather than spilling the beans like I did.

Also, this is my first post, so please extend some grace if I break any conventions! 🙏

welcome to embedded :)
one of the best resources online is this one:
<a href="https://github.com/m3y54m/Embedded-Engineering-Roadmap" target="_blank" rel="noopener nofollow">https://github.com/m3y54m/Embedded-Engineering-Roadmap</a>

Just as start… Start playing on Arduino and Raspberry pi , try to understand the electronics and electrical components and devices through experimental works

Getting into LVGL as a web dev was pretty easy. And they have a lot of examples for micro controller. But I already knew what a pointer is 🤣
I mean, yeah you’ll need some HAL level digging for a platform but it’s nothing compered to downloading 5gigs of npm archives to generate 100kb web page. Or making legacy C++ program to compile. C is a very simple language. It just pushes bytes between buffers. And it’s very debuggable unless you add concurrency. And the dreaded lack of GC isn’t that big of a deal since everything on stack is GCed.
It looks more like “getting out of embedded” to anything else is a tricky part. Probably the hardest thing is loosing the understanding of wtf is going on and accepting 1000 layers of abstraction as a reasonable approach.
Still can’t get into FPGA for video processing the tooling is horrendous vendor locked pipe of :poop: and your app is a one black box 😠