I passed 2 AWS Solution Architect exams in a month
bhaskargandavabi.medium.com
bhaskargandavabi.medium.com
Imagine if walmart was selling a certification on all the best way to distribute your money in their stores...
Anyway, everyone is free to do whatever he wants, but I'm very annoyed and worried about the network effect of such a thing.
For example, because of his bullshit aws certificates, this person will have the feeling of being part of a special club, and a sunk cost incentive to push for whatever possible aws solution to any problem instead of anything else.
And sometimes companies could use that as a requirement or plus in recrutement against other candidates, when, in fact, anyone will be able to do the same job by just spending a few hours only reading the doc for what is really needed!
I haven't met very many, but this is basically the attiude I've got from most people with AWS training.
Edit: With good/experienced developers, it's usually not that bad, but with new developers it can easily turn them into architecture astronauts.
Here's my original comment:
Well, the flip side is. Do you have experience with AWS? Not really, but I've known Linux for years, have done stuff on Heroku, Digital Ocean, my own servers, hacked boxes and IBM Bluemix, but no AWS.
Sorry, we rejected you because you have no experience with AWS.
So to be honest, I've been looking into it, because apparently a lot of Dutch (not all) are too short-sighted to see that I've done development and am able to figure it out, or have transferable knowledge (fun fact/example: in part via HN articles that showed how to load balancers [0]).
Also on the money side, it's not that bad. It's $150 for an exam. I think by reading docs and watching this video [1]. It'll be possible to get.
Everything I mentioned here is also true for pentesting as I've experienced something similar there.
Now that you've heard me complain, I am actually for hire ;-)
I'm language agnostic, most experience with NodeJS/ReactJS, junior pentesting as second skill, I'm at my best as a 50% programmer / 50% anything else <-- pentester may be included in the second 50%, or: marketing, business analyst, UX, I'm open-minded and willing to learn.
If no company is hiring me by the first of January and all my applications are dead in the water, I'm starting something for myself.
[0] https://news.ycombinator.com/item?id=21490731
[1] https://www.youtube.com/watch?v=RrKRN9zRBWs&t=32838s&ab_chan...
A lot of the initial challenge in interviewing is proving you can meet a certain bar to get your foot in the door for an interview. And $150 to prove that you can do a 100K+ /year job seems like a pretty rational investment
Anyone who comes to any cloud provider with that mentality actually is worse than someone who has no experience.
That’s just as bad as someone who models a schema for a database meant for analytics the same way that model a database meant for transactions and wonders why it isn’t performant.
I have checked their huge product offerings a few times when I needed something I know they must have, and used that, like AWS Batch and AWS SQS.
As for "all of the different services". No, I haven't, but I can, and I will do that, on demand. My point is exactly that I don't want to sink huge time into reading their documentations for something I may never use. And I don't believe it is correct to prefer candidates who did that.
But the attitude that a cloud provider is just a bunch of VMs leads to crap like hosting everything on VMs, reinventing the wheel etc.
You wouldn’t hire a database admin with no database experience and expect them to learn on the fly. Why hire someone with no experience with your cloud provider and hope they would learn on the fly?
Any technology/service/language has footguns and best practices that you only learn from experience.
We rarely (if ever) actually communicate our AWS certifications, except towards AWS themselves.
I got certs because the company wants them as part of their AWS contract. Other than that, the cert means nothing at my company.
1921: "Why learn a Underwood typewriter and learn all their marketing junk?"
It just actually means you want to learn how to type. Sure the brand you learn on is a "Underwood" - but the concepts and structure are transferable. YES! You don't have a CAPS lock feature.
Cloud companies are just as weird and envious as the rest of us. So they copy!! If you can learn s3 you wont be shocked by any principals of gcs. Cloud tech is transferable. Cloud containers? Docker works on both.
People tend to use binary logic to gloss over the details that determine our future.
K8s is opensource org backed, docker is opensource startup backed. sshing in and installing Apache is an open setup that can't be sabotaged but not open standard. Signing in to Linux and starting your own python or C++ server code is open standard and can't be completely hijacked even by a BDFL controlling the main implementation direction.
Give the content a try, you'll find at the foundation of it is building highly scalable distributed system, just that the execution of it is via AWS. You can apply the same knowledge to any cloud platform.
I doubt that. Looking through the link and recently having discussed this in an OT-thread of an hobbyist-forum, the only thing you learn is AWS-lego. I don't see, how training for marketing keyword-riddled multiple choice-questions helps you in understanding the systems in question (case in point: Q1 of the "blog" - basically the only skill required is basic (!) logic matching requirements to the services (my mother can do this) and translating the marketing speech to requirements... – you don't need to know anything about interdependencies.)
If certification is indeed useless, and exams are simply basic marketing exams, it will eventually become a red flag in the hiring process.
Gaming the system... maybe, but it pure profit.
It is very funny when aws came up with their own.
I think this happens when a company engineers something so complex you need to train people how to use it. Rephrased as when a company fails to make their products easily understandable.
Having said this there is a place for such complexity and with market share comes the right to do so.
The aws product line is quite comprehensive and they have broken new ground trying to offer this.
Of course if this was a google company I would imagine product creation and sunsets. Meaning they shouldn’t be doing this.
This serves as an artificial barrier to hiring. I've done years of network engineering but then it comes down to "how many years of experience in AWS."
EDIT: Note that this isn't a new thing, e.g. we're looking for an "Oracle engineer"/"vmware specialist"/etc.
A Raspberry Pi home automation server running Alpine, a load-balanced OpenBSD router... I make some projects which are considered a hobby for now.
But how would you convey this to a recruiter with limited to no understanding of technology? How would it even look like to a recruiter if someone like me was trying to get a foot in the door?
This is where certs might come in handy. I may not become an expert overnight, but at least it shows that I'm willing to focus and learn the material required to pass a cert.
For example, if employers would not find any good candidate with the certificate, they would not think of it as valuable.
Also, on the general topic, what should be more valuable:
A) an aws certificate on Route53
B) a certicate of training in dns and dns configuration
Which one will give you transferable knowledge to the other? But which one will be favored by a clueless HR recruiter?
This may be true for the lower levels (haven't taken those). But if you go higher, it's pretty heavy technical content. On the second level, I'd estimate <10% of questions/content was marketing.
Calling it purely marketing content is cynical because even documentation or tech talks could be interpreted as marketing if you are cynical enough.
You can always choose not to go through their hoops and just study learning content but not book an exam. Learning platforms don’t really care.
I'm fed up with people in tech thinking that certifications are a bad thing, or somehow a negative sign in a candidate.
The reason why is bc there are so many people in tech who are self taught. So when a senior self taught person without a cert sees any mention of a cert in their domain, they get insecure bc they do not have that cert. That senior thinks - "great, more stuff to learn." Then that senior vents to anyone, even on hacker news, whining that certs aren't necessary and people who have them are somehow less knowledgeable.
> I think this is not the type of person I would want to hire for AWS. It’s kind of like learning AWS for the sake of passing a course.
LOL WUT? How about that's exactly the type of person you want to hire. And you wonder why the knowledge levels between candidates vary so much and why there is a skills shortage.
How about - certifications are a good thing. It's a good thing that people with no knowledge have a goal of at least some level knowledge that someone more experienced somewhere decided that the level of which to pass such certification is a generally agreed upon base line.
SMH. This attitude needs to change, but I doubt it ever will.
Edit - yesssss, here come the downvotes. Keep em coming, boys.
My employer sponsored my oracle java 8 cert. At the time, I knew zero java, and I was able to convince my boss' boss that it would be a good idea after my boss denied and said it was useless (bc he did not have it and did not want to validate it).
Going through our code base after was eye opening. We were doing a few things wrong or in a non-standard way.
Basically everyone has the experiencing how little junior devs know right out of college but don't seem to want to apply this logic to certs. "I took a class on AWS in college" is about the equivalent of a cert and if that's the best you can put on your resume it doesn't look great.
When I am hiring someone and I see they have certifications that means they put in extra effort. It’s not a substitute for experience by any means but there is some foundational knowledge that comes with it.
The certifications are not easy and do require studying and actual experience in the AWS ecosystem.
If a team wants to look down their noses at someone who has a certification then that is a red flag to me as a potential hire. I probably wouldn’t want to work with people like that anyway.
I will continue to put certifications on my resume with pride and if they don’t like it that is on them. I’ll be fine regardless of some unnecessary elitism.
As very anecdotally, I know one person with an Amazon cert and they are also someone who wants me to tell them which line numbers of their code to edit for assignments. And they got it by cramming for the exam like any other test. So my thoughts on Developer Associate are not very high.
Other developers have told me that one of the easiest ways to get a cert is to just pay someone in India to take the test for you in case you need it for a contract (for those developers in the consulting world).
Both being able to cram for it and it being required as part of crappy contracts but not genuinely mattering if the people doing the work did it seem like very negative signals about its value.
That would depend on the cert and the level of experience of a given candidate.
> As very anecdotally, I know one person with an Amazon cert and they are also someone who wants me to tell them which line numbers of their code to edit for assignments. And they got it by cramming for the exam like any other test. So my thoughts on Developer Associate are not very high.
Obviously, that is more of a reflection on that person's personality and less on them having a cert. Not everyone with a cert is like that.
> Other developers have told me that one of the easiest ways to get a cert is to just pay someone in India to take the test for you in case you need it for a contract (for those developers in the consulting world)
That's a culture problem, not a cert problem. Some people get certs to honestly learn and signal that knowledge to others, not just to tick a check box or add a resume bullet point.
Yeesh. SMH.
But certs cannot be isolated from culture, which undermines their value.
The whole point of any cert is to serve as a proxy for knowing that someone else knows something. You aren't a doctor, and have no way to judge anyone else's fitness to prescribe or perform treatments on you, but they have a diploma that you are assured by everyone you ever met that those are hard to get, especially if you look one level deeper and look at the school name rather than just the fact of a diploma at all.
If you can be useless and still get the cert, then the cert is not only of no positive value, it is of actively negative value, at least in the macro scale for everyone as a whole.
It can obviously be of positive value to an individual as a game move within a disfunctional game.
But that just comes at the expense of everyone else, like capitolizing on some tax loophole or something instead of seeking to close it.
A couple of reasons:
1. It's a weak anecdote 2. Anecdotes are not statistically proven facts 3. I have anecdotes from my life that demonstrates the exact opposite
I do not know how you read an anecdote, on the internet of all places, and take it seriously.
I am a consultant at AWS, I have eight (soon nine) certifications I work with many industry experts. I would say they aren’t good at all at assessing someone’s skill. At most if you take them seriously and go beyond the crash courses they help you be conversant on the services at the high level and on areas that you already know they may help fill in knowledge gaps. My entire purpose of getting the certs was to know what s
For every single service I’ve had to use with AWS, I’ve had to dig into the documentation, the APIs, blog posts, Udemy courses for AWS’s managed versions of open source software like ElasticSearch, etc.
It’s way too easy to get certifications without any experience.
Nobody is against training and that's why companies still encourage juniors to take training and even some certifications. But some certifications aren't really designed to be a good training. Just go on internet and see how people are seeking for tips and tricks to handle the exam itself and also check study materials and you will understand where that opinion is coming from.
You actually need practical experience with using any of their services in production, to figure out capabilities/limitations/warts/etc.
I learned through LinuxAcademy.com (dunno if it exists in the same form anymore).
I was hired to help a group of linux admin seniors who knew a ton about their work, but had pretty terrible AWS knowledge. After 2 weeks (!) I was able to improve architecture of our environments enough that I made thousands in savings per month and make our product more resilient at the same time.
All this gain just because they hired someone who learned how and why to so things through a good cert course. Instead of using the platform as a pure KVM alternative I showed them the APIs that make things easier for us.
I have eight and soon nine AWS certifications I’ve only ever paid for one out of pocket. The companies I worked for paid for the rest. But I would say that the certifications prove nothing as far whether you are proficient in the certain skillset.
I got my first certification without ever logging into the console. The only reason I got any of the certs (besides the 7th one that was required when I was hired at AWS ProServe) was to have a guided learning path with a goal at the end. The most they do is let you know what you don’t know and the “what” but not the “how”. I am consistently disappointed about how easy the certifications are compared to what I really need to know on the job.
How can you get a developers certification from AWS and not know the first thing about development? How can you get a database certification and not know any of the query languages involved? I needed to learn about ElasticSearch a couple of years ago for a project. I bought a course from Udemy that was 20 hours long to learn the ends and outs of it and they even admitted glossing over some advanced parts.
You can take a 14 hour course in AWS Data Analytics and pass the course. The exam consists of Redshift (an OLAP database), ElasticSearch, various Apache frameworks (Hive, Presto, Spark, Hadoop, HBase, HUE, etc.), Kinesis (stream processing), Glue (Serverless PySpark infrastructure) and a few other services. Do you think that certification would qualify anyone to actually implement any project on AWS? Should I have hired someone based on the certification when I was in the real world?
While I disagree with all of the posters who say stuff like because the used Digital Ocean for their Todo app that they “know AWS”, I would hire a developer with no AWS experience before I would hire someone with an AWS Developers certification with no development experience. The same is true for all of the other specialities - Databases, Data Analytics, ML/AI, etc.
Except for infrastructure and networking. I’ve seen far too many on prem network folks who got one certification and called themselves “consultants” and ended up costing the company more because they didn’t know how to leverage services being offered.
But... I fully realize I'm doing this for fun. I wouldn't claim to anybody it actually makes me good at doing any of the things I'm "certified" in. So while I'm confident the author's approach is efficient and effective if your sole goal is "pass exam, obtain cert" - if there's any practical goal after that (e.g. "get employed and be good at doing this stuff"), I doubt this is a good approach. Maybe a good starter, maybe even not.
I'd be curious what's HN perspective?
It's a signal among other equally valid signals and while it doesn't replace actual experience, it's still a nice plus.
I'm less cynical about certs offered by neutral third parties, or if they are offered free of charge. (If quality versions of those exist?)
I am trying to fit this message on a tshirt. Any input would be greatly appreciated!!!
Idealistically the knowledge is so transient that it’s probably only worth googling on the fly. I’d rather fill my head with mathematics or something instead if I had the choice.
This I learned without certification that were more valuable: C, Go, Linux shell and admin, HTML, CSS, networking, various reusable computer science bits.
You still need experience, practice and hands-on / real-world experience though, to be actually good at it I feel.
Maybe most people don’t need to know. But the truth is that a lot of these managed cloud services are wrappers/reimplementations around OSS tech or can be a massive overhead for your actual problem.
A great exercise to figure this out is to implement some of the cloud products as a toy project, targeted at your use case. You’ll quickly learn whats fluff and where the meat is.
Or what about Go? React?
Ok, some of these aren't proprietary, that if the big company supporting it would drop?
My point isn't to now learn AWS because everything else is proprietary. My point is that quite a few popular programming languages have corporate interests and can also give a bit of a vendor lock-in, to quite a degree.
I think the bus factor for all of these technologies is not the same. Nor is the ease of switching. If AWS died tomorrow, people would switch to Google Cloud. If Facebook dropped React, it's insanely likely someone would come and take over (esp. since it's already open source).
The cost in switching technology is just too big for big libraries/big languages (e.g. even with Oracle somewhat dropping Java, it's still top 3 of must used languages).
C# for servers seems at the same point nowadays (and before you could still own software and hardware a lot more than today).
Go - problematic, but GCC has a compiler, which can bootstrap.
React - now, we are talking about the mess, the www is today, but technically you'll be able to compile webkit for years to come. And it has some big stakeholders.
AWS - everything is proprietary you can't "own" any tool at all. Amazon decides they don't want to do business with you - pooof. Your country decides to block AWS/USA blocks your country - pooof. Want to fix a bug - well, nothing you can do...
The real skill of AWS is knowing the cheapest way to use it, not the easiest.
> A company currently uses Amazon EBS and Amazon RDS for storage purposes. The company intends to use a pilot light approach for disaster recovery in a different AWS Region. The company has an RTO of 6 hours and an RPO of 24 hours.
> Which solution would achieve the requirements with MINIMAL cost?
You have 4 multiple choice questions, however only one uses the "Use EBS and RDS cross-region snapshot copy capability" while the other uses generic server jobs. It is therefore obvious which choice is the right one without even knowing anything about AWS, all you need to know is that they want you to pick the most specific option they have instead of implementing your own services.
And most questsions were taken directly out of Oracles marketing material. This was in 2010 but one of the questions was like "Why is it much cheaper for a client to use Oracle xyz.." and every answer was like a pat on the back of Oracle. Felt so dirty afterwards but I passed the exam ;D
For this question you for example need to know that such cross-region snapshot functionality exists. It could also be an answer made to confuse you, because maybe there is only same-region snapshot functionality at the moment. Also conditions like "minimal cost" can make an obvious looking answer the wrong one if don't know details about the cost of the affected services.
Do they put services AWS doesn't have in the answers? I'm pretty sure they wouldn't since it would make themselves look bad. But yeah if they put in a lot of nonsense answers I agree it works to prove that the guy knows AWS, but I wont believe they do until someone proves otherwise.
> Also conditions like "minimal cost" can make an obvious looking answer the wrong one if don't know details about the cost of the affected services.
Not really, the only time you can make sweeping statements about "minimum cost" is when it is both minimal both in engineering hours and running costs. Therefore the solution with the least number of engineering hours spent will be the right answer.
If your company wants to participate as an AWS Solution Provider, there are hard requirements for the number of staff with AWS certifications. Most Fortune 500s will only work with AWS Solution Providers, so there is a huge incentive for consulting companies to encourage or even require their employees to become certified. This can involve holding back promotions, paying for test fees, allowing the employee to study on company time, bonuses and any number of "carrots".
Also, AWS certifications expire in 3 years from the date of the test pass which I think is a good timeframe, given the historical rate of change.
I realize that this is basically the same relationship that MSPs have but at a certain point isn't this just "working for Amazon" with extra steps?
I think they are just emulating the successful Microsoft VAR model that had different tiers from Microsoft Certified Professional to Microsoft Certified Systems Engineer to Microsoft Gold partners. Microsoft probably just imitated Novell and Cisco. The business model works. I don’t think Cloud adoption would be anywhere near as high without corporate-approved partner designations with multiple tiers.
Now seems like a good time to pause and take the time to do some certs.
- Instead, learn some scripting (bash, python). You will need to constantly work on improving this.
- Learn how to set up a Linux server and get a website going, from scratch. Even if you won't do it much in real life, it's helpful to know how it's done.
- Get at least AWS Associate cert.
- Find an employer/manager who will let you work with AWS in lab/production in real life.
But I don't like them. They are very broad and often ask quesions about legacy services. You don't need to know all the intrinsics of EC2/VPCs if all you use is Lambda.
If you want to get a good overview over AWS tech, take the associate exams.
If you want to learn every detail about AWS, take the speciality and professional exams.
But if you wanna build something good quickly, your time and money is better invested in courses focusing on specific problems or technologies.
There are also some questions that are just wrong because they haven't been updated. Definitely found myself thinking, "ok, technically this is now correct but that's only been true for a few months; this question probably actually wants this other answer that used to be true."
I don't remember a thing about all others because I don't use them at all anywhere. They look good on the CV but that's about it.