If you are not using AI to write the
Saved transcript
What Every .NET Developer Actually Needs to Know in 2026
Channel: Milan Jovanović
majority of your code in 2026, you are
risking seriously falling behind as a
developer is something I would say if I
wanted to FOMO you into watching this
video. However, I will share my
practical .NET developer road map for
2026 where I won't be focusing on any
given library that you should master
because these are mostly implementation
details. Instead, we're going to focus
on the fundamentals. I'm going to break
this down into a couple of highle topics
that I think you should have mastery of
to be a proficient .NET developer in
2026. Now, I'm calling this the 2026.NET
road map fundamentals edition because I
don't want to focus on libraries. These
come and go pretty often. Some of them
that are popular today go commercial
tomorrow and people just stop using them
and move on to the next thing. So,
that's why we have to focus on the
things that don't change. And number
one, this is going to be net and C. So
in this section, I'm going to say you
want to strive to be on .NET 10 or at
the very least .NET 8 or 9 as these are
the latest versions, .NET 10 being the
long-term support release. They have the
most performance improvements. They also
have the latest developments and APIs in
the various libraries that you would be
using. So that's what I would be
focusing on if I was looking to upskill
myself as a .NET developer. Now when it
comes to frameworks, we have ASP.NET
Core. This is still your bread and
butter for the majority of net
applications out there and I'm going to
say that you probably want to focus more
on minimal APIs but you should also know
controllers as these are going to be
transferable between the many popular
frameworks if you want to for some
reason branch out of net another key
concept is the or dependency injection
you should understand how this works
both from an implementation perspective
between the various lifetime scopes that
we have in ASP.NET NET core as well as a
design concept for building your
applications and what benefits you get
with dependency injection. Then you
should master off and security. We have
a few options for this inn net and you
should be spending some time learning
how you can implement this. And then
lastly, I'm going to flag testing with a
focus on integration testing preferably
using something like test containers as
this allows you to mimic your complete
development environment to have maximum
confidence in your tests. If you master
these things, you're probably 90% of the
way there of the things you will need to
be proficient as a .NET dev. The next
thing you're going to need is some sort
of database. And I'm going to give you
two choices here. You choose whatever
works best for you. It's either going to
be poss.
I know some of you are expecting
something exotic like a NoSQL database.
But what I'm focusing on is something
more fundamental than the specific
flavor of SQL that I have here. And that
is the underlying concept that is SQL or
relational databases. If you understand
SQL and how it works, you'll be able to
apply these skills in any kind of
database. My personal choice is posgress
because number one, it's awesome. Number
two, it's free to use and it's supported
in most cloud providers. And you can
build pretty much any type of
application you can think of using
Postgress. But more on that in a moment.
I wanted to highlight SQL because there
are a couple of concepts here that you
should understand. So these are going to
be data modeling and how to structure
your database schema. Then indexing to
be able to improve the performance of
your application queries. And this one
might be a bit on the advanced side, but
it's understanding query plans using
something like explain analyze in
postgris. Now I actually did a video
about this quite recently. I'm going to
link the video in the description and
it's also going to pop up in the corner
of this video as you're watching.
Understanding all of these things will
help you use your SQL databases to their
maximum potential. Now, let me explain
why I favor posgress. Posgress isn't
just a relational database. It's
actually an object relational database
and this means that you can do some
interesting things with it other than
justing rows and tables. So,
fundamentally it is a relational
database, but it also has very extensive
support for JSON. And there are
libraries out there that will allow you
to treat Postgress as a makeshift
document database. Now, it also has a
rich set of extensions in the ecosystem.
For example, something like PG vector.
And suddenly you can turn your
relational database into a vector
database and start implementing some
advanced features like semantic search.
It also comes with a bunch of other
extensions like PGRON that lets you do
background jobs inside of your database.
You can even efficiently store time
series with something like time scale DB
and there's lots of samples out there
that covers how to implement all of
this. As I said, you can run posgress in
any major cloud provider. So whatever
you choose is probably going to work.
Then number three, I'm going to say you
need mastery of messaging to some extent
at least because most modern
applications are going to grow out of a
single server and a database and start
needing to do some more asynchronous
work and this is where messaging can
come in. Now, you could definitely use
Postgress to also cover this part, but I
probably recommend using something more
standard here. And I'm going to give you
a couple of suggestions. My default
choice if self-hosting my message broker
is Rabbit MQ, but I also recommend
exploring Azure Service Bus or if you're
using AWS, you have SQS and SNS to be
able to implement the majority of the
messaging patterns that you might need.
Now again we have an underlying concept
that we should understand when it comes
to messaging and these are the various
broker topologies between cues and
topics. Cues allow you to send a message
on one end and consume it on the other
and only one consumer can pick up the
message. Topics allow you to implement
broadcasting and a couple of more other
patterns. Then you need to understand
how item potency works. And this is very
important when it comes to messaging
because you have to think about things
like delivery semantics with the default
one being at least once, which means you
may get a message delivered once or more
times. And this is where dduplication on
the publisher side and item potency on
the consumer side come in. And I covered
both of these topics in a recent video
where I was talking about the item
potent consumer pattern. If you're
looking for something more advanced, you
can explore the outbox and inbox
pattern. But nonetheless, this is going
to be a fundamental topic that you need
to master regardless of which library
you are using to facilitate actually
sending messages from your .NET code. So
with these three things, you can
probably build like 90% of the
applications out there and if you need
anything extra, you can just figure it
out on the go. Now, when it comes to the
cloud, this is definitely important. And
here I'm going to keep it very simple.
I'll give you two options. Either choose
Azure or AWS. You can just pick one. Any
one of them works just fine. If you
don't know how, find a getting started
guide out there and figure out how to
build a simple application and get it
deployed to the cloud. So that's the
next thing when it comes to the cloud.
Learn to deploy your net applications
and preferably you want to figure out
how to do this using some sort of
continuous integration system. The
simplest one probably being GitHub
actions. you can build most of the
patterns you need using it. So I think
this is a good option to get started.
But really whichever cloud provider you
use, just figure out how to deploy your
application. This is going to help you
stand out from the crowd so so much you
may not even be aware of it. And then
number five, the hot topic which is AI.
And I'm going to actually scope this to
AI tooling which I think you should
definitely be using if you care about
your time and you want to be an
efficient developer in 2026. Now when it
comes to the specific tool that you are
using, it doesn't really matter. Just
pick one you like. Let's say for IDs I
use cursor and if I wanted to do CLI,
you can use something like clot code. I
think these are the most popular options
with my preference probably leaning
towards cursor and using an IDE. And if
you think that AI only generates slop
code is probably a skill issue and you
don't know how to properly use it. To
give you a quick example of using AI,
here's how you can use cursor to
implement something that touches most of
your application like let's say taking
all of the query handlers which are
using EF core here and let's say we want
to convert this into raw SQL queries to
make it more low-level and faster. So
what you could do is open up each query
handler and rewrite the queries yourself
or you can tell cursor and this is
important to plan out this migration.
So, I'm going to use the plan mode to
kick this off. And this is going to scan
my codebase and produce a markdown
document outlining the steps that are
required to build this feature. It may
also ask some clarifying questions such
as how to obtain the database access.
I'll say that it should create a new
abstraction. I don't want to migrate my
commands, only the queries. And I'll
press continue. And then it's going to
produce this document which you can
treat as documentation and store it in
your repository. And it's going to
outline how it wants to implement this
feature. And this is where you come in.
You can add more inputs and steer it in
the direction that you want to. And then
when you're happy with everything, you
can allow it to implement this feature,
analyze the codebase, and then it
becomes very valuable. If you have an
extensive test suite, you can run your
tests, make sure that everything works
the same as before. And from there, you
can go ahead and make a pull request and
continue iterating on this feature if
required. So instead of just treating AI
tooling as your VIP coding buddy,
actually try to approach this from an
engineering perspective. Come up with a
requirements document that you can hand
off to your AI agent to build out the
feature that you want. Most of them also
support various rule files where you can
define your code styles and any
preferences that you want to enforce and
a lot of these things can be automated.
Uh when it comes to my personal
preferences for each of these sections,
it's probably going to be posgress over
SQL server, rabbit MQ if I am
self-hosting. Even though I very much
enjoy AWS, I will still choose Azure for
hosting a .NET app. And then I already
mentioned this, I like using cursor for
my AI development, but you could also
use many other things. For example, VS
Code Copilot or even Visual Studio
Copilot or I don't know if Chatra Writer
also has built-in AI tooling. It
probably does. I just don't know what
it's called. And then here's a bonus
section if you really want to stand out.
So let me name this as bonus. And here
I'm going to list out a couple of things
if you want to go beyond just being a
.NET dev and into full stack developer
territory. And I have to say that with
AI tooling, there's really no reason for
you not to be a fullstack dev. I'm not a
fan of the .NET UI frameworks. So I
recommend using either Angular or React.
I used to be more of an Angular fan.
That's all I knew back in the day. Then
I learned React. And basically all of
these are just component-based single
page applications that once you figure
out how to use one framework, you can
use any one of them proficiently. And
then when it comes to the language, I
favor TypeScript over JavaScript.
Although you can definitely start with
just JavaScript and then figure out how
to use this in TypeScript. So that's my
fundamentals road map. I know it's not
exactly a step-by-step instruction for
how to become a better net developer in
2026. But honestly, you should just pick
somewhere where you don't have as
proficient skills. Find a side project
or application that you can build and
then just go out there and find some
time to build this application.
Preferably, it's going to be a full
stack application with a UI and the back
end. You definitely want the database.
You can consider adding a cache in
there, although it's not required as
much. Messaging can also be optional,
but I highly recommend learning how to
get this deployed and how to automate
these deployments using a CI tool like
GitHub actions. And then you definitely
want to learn how to be proficient with
AI tooling to make your development
faster. This is also my first attempt at
making some sort of road map. So
depending on the feedback, I may do a
more in-depth video in the future, and
I'll probably go back to this in a year
from now to see if and how my
recommendations might change. If you
want to grab this fundamentals road map,
it's going to be available from the
pinned comment right under this video.
And if you're looking for what to watch
next to improve your skills in 2026,
then I recommend checking out this
video. Consider smashing the like button
if you enjoyed this. Thanks a lot for
watching and until next time, stay
awesome.