IT/Software career thread: Invert binary trees for dollars.

Khane said:

How expensive were the Salesforce/Snowflake certs/training?

Click to expand...

$200-$400.

The Snowflake Data Engineer one was $375. Work comps me for all of them so its whatever. I don't bother with official trainings. Just read documentation and go through practice tests is the tried and true method for me.

Resources I have used:
1. Udemy courses (Chris Garcia ones for Snowflake are spot on).
2. https://focusonforce.com/ for Salesforce.

EDIT: Salesforce is especially appealing because of the platform structure. Anyone can spin up an environment and just test stuff out and they have loads of tools to support doing that. The VSCode Salesforce extension (there are several) can have you write code or configure stuff right in the IDE and sync it to the environment.

Snowflake might but I've never not had access to the enterprise version of it so I am not 100%. This can really prevent you from understanding some of the docs and behavior as you are unable to actually do it yourself and see how it functions.

Jokes aside about "Vibe Coding" the interview that started the viral term is going to be prophetic. Within a few years you will have entire [software] businesses running on teams of "vibe coders" who sit there and chat with a LLM like Cursor AI and produce working code. Code that they do not understand in the slightest and cannot even read without a summarized AI prompt explaining to them what a bunch of code lines are doing.

It's going to be some wild shit bros. Might be a good opportunity for us oldbros who actually understand computers and programming. Because somewhere along the line the results of mass vibe coding will be apocalyptic.

More like, the model isn’t quite getting this right please fix it?

Deathwing said:

Are we going to be this generation's COBOL programmers that banks need to hire to maintain their software?

Click to expand...

In a way.

The next generation will be born into these abstractions. They may be "developers" and the application may be in Java but their interaction with it is going to be talking through an LLM to generate it. They will not understand a lick of it. If the LLM isn't able to get it right or does something that is less than optimal and causing downstream issues they will not be able to diagnose and override this. Say, some issue with memory management or something. They will not grasp the principle of this. Only that they need to keep talking to the LLM until it goes away.

Things that are poorly documented (lol) or not documented at all (lol lol lol) are going to be major gaps in the LLM's understanding. Especially in proprietary technology that does not have far reaching support. It will be on companies to train internal AI agents enough that they can work with it on a deep level.

Companies will be founded by people who have never written code proper. What I really mean by this is someone who never really had to read through function documentation, or write their own. They just use whatever the LLM spits out.

Neranja said:

The industry needs more programmers, because management sees this a competitive advantage, as more and more business logic is depending on IT: You have to automate your business or you will die.

But competent programmers are a rarity nowadays. Industry tried everything to get new programmers, like SCRUM, code interviews, certificates, education, dissipating responsibility (e.g. microservices). Nothing seems to work, as code has become so complex that maintaining it is even more of a shitshow that it has been 10 years ago. Even throwing a horde of mediocre to bad coders (the famous Indian code shop) doesn't work anymore, so now they are trying the "AI will save us" card.

Do you want to feel really old? Story time: Some years ago we had an 18 year old recruit who started his education in my team, and even with his high computer affinity he didn't know what a partition and filesystem was. He didn't last three weeks.

Click to expand...

Speaking of.

I consider myself a humble person. I know I have a strong work ethic that I relate more along the lines of being far too dumb to quit and also being maybe a touch obsessive when I get into some issue. Especially when it comes to problem solving. I am sure none of you can relate! I do not consider myself any kind of super genius but I am at least competent in my specialty. That being all aspects of data infrastructure. If it is not SQL, Python, or Java I consider myself functionally retarded. More than usual. I've automated a lot of stuff here and built out a lot of the infra in our org. Then people tell me I'm super smart and intimidating. Like a coworker told me this directly. I didn't even know what to say.

I have been doing technical interviews this last week for our data science team. As the manager desperately needs someone she does not need to teach the programming part to. The product data is of enough complexity that a new hire will take at minimum six months to understand it enough to reliably aid in the algorithms she is tasked with engineering. These algos are all about predictive product use and whatnot. So the technical interview is a strict requirement. They need to pass it to be hired.

I wrote 6 questions. 3 for basic SQL and 3 for basic python wrt data science. Straightforward data manipulation with Pandas that is just about as standard as it gets. I interviewed two fresh CS graduates from UT Austin. Both had very clearly never written a line of SQL in their lives and knew nothing about Pandas. But could write some python. The third was a non-CS guy with experience as a data analyst. He knew significantly less than the other two and tried to bullshit his way out of the technical interview as he thought there wouldn't be one (???).

The quality of technical people is surely going down as these kids were all born in 2003 and stuff. Damn we old.

Khane said:

They graduated with CS degrees and don't know any SQL whatsoever? How is that even possible?

Click to expand...

Python is the lingua franca of CS programs these days. Any interaction with SQL is extraordinarily limited. What they knew of it was seriously limited to python sql lite. They had no idea what MSSQL or MySQL was. Their knowledge of Java or C is also very limited. CS is a very watered down experience where you can avoid lots of the harder principles of computers and programming. At least from UT Austin and other schools in the state.

Khane said:

I would say I'm surprised but I guess I'm not. The lack of talent of any new hires at my company since like ~2018 has been so bad it's almost comical.

Click to expand...

Here are two of the questions I wrote. I spent a whole 15 minutes writing the 6 code question bank for the interview.

Not one of them was able to answer this question. This is an entry level position and I consider this basic bitch shit.



Or this one.

TJT said:

Here are two of the questions I wrote. I spent a whole 15 minutes writing the 6 code question bank for the interview.

Not one of them was able to answer this question. This is an entry level position and I consider this basic bitch shit.

View attachment 581068

Or this one.
View attachment 581069

Click to expand...

Man, Im a huge noob hack and even I can do the SQL, and could probably get lucky with the Python.