Career Guidance in AI/Data Science for a Student Who Appeared for IGCSE 10th Exams in March 2025

[Futureized]

Exploring Career Pathways in AI/Data Science After IGCSE (PCM/ICT/English)
Student has recently appeared for the IGCSE 10th exams in March 2025 from Mumbai, with subjects including Physics, Chemistry, Mathematics (PCM), ICT, and English. Considering a career in Artificial Intelligence or Data Science, we are currently evaluating the following post-IGCSE options:

1. JEE + Integrated College Program
2. MHT-CET + Integrated College Program
3. Post-12th: 4-Year B.Tech Program at Atlas University in Mumbai
4. Cambridge A Levels (11th and 12th)
   Note: A Levels appear most beneficial for students planning to pursue higher education abroad after 12th, based on current research and expert opinions.

Query:
Beyond these mainstream pathways, are there any additional viable options available for students with this academic background?

From the research conducted so far, the majority of students in similar situations seem to converge on two main routes: JEE or MHT-CET.

Given the long-term goal of building a career in AI/Data Science, what would be your best suggestions?

Also, if you’re aware of other students who are pursuing similar goals or considering alternative routes, their examples or guidance would be greatly appreciated.

 

[drumstick]

Hello i am a igcse board student i just finished my as levels but i have seen many of my friends suffer while studying for jee as they did not have enough time 2 years even other juniors who have been studying since grade 9 find it hard due to totally different syllabus and no calculator etc. so i would suggest trying sat exams and trying some uni which prefer sat exams. Also it would be better if they were to start learning the basics of computer science in as level along as it is much better that ict

 

[altair21]

Unless you want to be a "glorified" data muncher, there's zero scope in India for any core AI job and this is true for other countries as well, if you really want to get into core AI (which I doubt you want to), A-Levels/SATs/GRE is the way forward, if you can afford it, do well in SATs and build your career in USA, or if money is an issue, do B.Tech here, and then a MS and PHD from America/Europe.

I can expound more on it but I need more details on what exactly you are envisioning, because you might as well say I want to work with computers and build a career out of it for all the good it does. AI/Data Science are just buzz words thrown around willy nilly for non-tech folks.

 

[Futureized]

SATs/GRE is the way forward

Can you provide more information about this.

And any review for a B.Tech in AI from https://atlasuniversity.edu.in/

 

[altair21]

Can you provide more information about this.

SATs are kind of like Mains but for colleges in USA, similar process as in JEE but its more of a general aptitude test instead of something like JEE Advanced, you need a valid SAT score to apply for colleges in USA for B.S. (Bachelor of Science), GRE is similar to SATs but for graduate programs and is valid across US, Canada and some countries in Europe (not all afaik).

And any review for a B.Tech in AI from https://atlasuniversity.edu.in/

I wont suggest any AI crap degree from even IITs, specially from a no name not even deemed (from a cursory google search) university, all such universities will just scam you out of your money. Get a proper CS degree from a decent established private college like SRM/VIT (yes, these are decent) which have proven track record of placements, or at least have companies visiting their campuses (this honestly should be your #1 benchmark for any college you consider), even if you graduate from IITK/IITD/IISC, you will not be scoring any core AI job unless you consider building mind numbing statistical models AI, you need a MS/PHD for orgs to even consider you, Applied AI which is most likely what you are gunning for anyways is easier, but these days it mostly consists of LLM Prompting, maybe some embeddings, some trasfer learning, and thats about it.

My point is, don't consider AI as your future field, get a general CS degree which will give you a wholistic base as well as exposure to other fields in CS, then decide what you want to do, I was like this before
AI even became a buzzword, wanted to work training my models, applying research papers to real life problems and what not but reality is a lot different.

 

[b.life]

Hi,
Recently post regarding data science was made and I replied there and don't want to repeat it once again so linking to it.
Regarding AI, Artificial Intelligence can be loosely classified into knowledge acquisition(using techniques like neural networks to convert raw data into usable information) and decision systems(using the knowledge to perform actions, like object identification, robot control, logic analysis, etc). But these days running machine learning code written by someone is AI.

The optimal approach to making a good carrier in either AI or Data Science field, is to take a generic CSE degree from some good college and on the side
1. For carrier in data science field,
1.1 Study statistics, and probability, business analysis, R among other things.
1.2 Identify which domain they like to work in. Like Business, Medicine, Finance, Political Analysis, Market Research, etc and identify the tools used and start specializing in them.
1.3 Do 1.2, from Year 1 of college till Year 3, and for Year 3 & 4, start writing reports from publicly available data and send it to interested parties(for eg, do a market survey about furniture build a good report and send it to reliance/tata these companies has research teams that does real data analysis works). Most wont reply but some will and you can build connections.
1.4 When you are finishing college, talk with the companies/think tanks that replied and build on the connection and you will get a good position based on previous reports.
1.5 Market research, Political and Strategic Planning think tanks pay a lot for good analyst. But the entry criteria is difficult but as the student is only now starting college they have 4 years to develop the skill.

2. For Carrier in Artificial Intelligence,
2.1 Be very through in calculus, algebra, statistics, and probability, skip this and you will have problems latter.
2.2 Learn LISP, Prolog, they are not the cutting edge or anything but when you program in them you will understand logic and flow better than in other languages.
2.3 As AI is a vast field, you need to decide which field you like for example, vision, speech identification, robotics among others.
2.4 Learn to code from the ground up and not use existing software. For example not many can write a 3 layer neural network from scratch.
2.5 Do 2.1 till 2.4 for Year 1 till 3. From Year 3, start contributing to blogs, opensource projects. Also when you improve an existing procedure try writing to the companies that use them and you will be surprised how many actually respond. Do not skip making your efforts visible else it will be very hard to get any traction within the AI industry.
2.6 Sad state but the truth is market intelligence(people profiling) is the highest paying job(next to Geo-poltical analysis because of government money) in the market right now. This combines AI and Data Science and pays are very good.

AI industry within India is almost non existent but Data Science in form of Market Intelligence, Financial Risk Assessment, Credit Card Fraud Detection, Think Tanks exist. These are not run of the mill job that get advertised but mostly filled in house or foreign experts brought in to setup and maintained in house(most of the banks does that). Reliance and TATA group haired data analyst to comb through their retail data some 6/7 years go but then they have gone quite, either they have established a good team or brought in foreign tie ups. Think Tanks access to quality analyst are shit and it shows in many of the reports that are published here, so that is something to think about.

I too will second the conclusions drawn by @altair21 on Indian colleges, so done waste money on specialized courses instead select a computer science and engineering course that has Theory of computations, AI, Neural Networks and you will get a good basic understanding that will help build up your knowledge later.

Regards

 

[altair21]

Hi,
Recently post regarding data science was made and I replied there and don't want to repeat it once again so linking to it.
Regarding AI, Artificial Intelligence can be loosely classified into knowledge acquisition(using techniques like neural networks to convert raw data into usable information) and decision systems(using the knowledge to perform actions, like object identification, robot control, logic analysis, etc). But these days running machine learning code written by someone is AI.

The optimal approach to making a good carrier in either AI or Data Science field, is to take a generic CSE degree from some good college and on the side
1. For carrier in data science field,
1.1 Study statistics, and probability, business analysis, R among other things.
1.2 Identify which domain they like to work in. Like Business, Medicine, Finance, Political Analysis, Market Research, etc and identify the tools used and start specializing in them.
1.3 Do 1.2, from Year 1 of college till Year 3, and for Year 3 & 4, start writing reports from publicly available data and send it to interested parties(for eg, do a market survey about furniture build a good report and send it to reliance/tata these companies has research teams that does real data analysis works). Most wont reply but some will and you can build connections.
1.4 When you are finishing college, talk with the companies/think tanks that replied and build on the connection and you will get a good position based on previous reports.
1.5 Market research, Political and Strategic Planning think tanks pay a lot for good analyst. But the entry criteria is difficult but as the student is only now starting college they have 4 years to develop the skill.

2. For Carrier in Artificial Intelligence,
2.1 Be very through in calculus, algebra, statistics, and probability, skip this and you will have problems latter.
2.2 Learn LISP, Prolog, they are not the cutting edge or anything but when you program in them you will understand logic and flow better than in other languages.

bruh whats this AI garbage that you are copy pasting? you definitely dont need either LISP or Prolog for AI, and if you really want to learn the extreme basics, then C/C++ is your best bet, Is prolog even being used anywhere? same with Lisp? atleast fact check before you copy paste shit

 

[b.life]

bruh whats this AI garbage that you are copy pasting? you definitely dont need either LISP or Prolog for AI, and if you really want to learn the extreme basics, then C/C++ is your best bet, Is prolog even being used anywhere? same with Lisp? atleast fact check before you copy paste shit

1. I don't know what you understood in this "Learn LISP, Prolog, they are not the cutting edge or anything but when you program in them you will understand logic and flow better than in other languages."
1.1 But I stand by my words, Prolog is the best for writing logic programs, if the user is new to writing complex logic, they can easily do so.
1.2 If you don't even know where LISP is used today then I have nothing more to add here.
2. If you want to have a productive discussion the first step is to stop insulting others.
3. Since you cant seem to differentiate AI generated text from others, I have nothing more to say.
4. Sorry to @Futureized for derailing the thread with reply to the quoted author.

 

[altair21]

1. I don't know what you understood in this "Learn LISP, Prolog, they are not the cutting edge or anything but when you program in them you will understand logic and flow better than in other languages."

what exactly do you mean by "complex" logic? like you do know that Lisp is on its deathbed and that prolog is dead? you might as well recommend fortran, heh or at this point why not just use assembly? since that will help build your "logic"

1.1 But I stand by my words, Prolog is the best for writing logic programs, if the user is new to writing complex logic, they can easily do so.

my point above

1.2 If you don't even know where LISP is used today then I have nothing more to add here.

please enlighten me, and please dont mention decades old infra surviving on nothing but prayers and patchwork, or old researchers who are used to it and just wont change, hell, name me just one new (started in the last 10 years) project thats using Lisp

2. If you want to have a productive discussion the first step is to stop insulting others.

and I would if it wasnt blatantly clear that you are just copy pasting stuff without doing any research, if I wanted that I would stick to shitty medium articles or reddit. I love TE because peeps post what they know instead of just copy pasting rando AI generated crap

3. Since you cant seem to differentiate AI generated text from others, I have nothing more to say.
4. Sorry to @Futureized for derailing the thread with reply to the quoted author.

 

[b.life]

what exactly do you mean by "complex" logic? like you do know that Lisp is on its deathbed and that prolog is dead? you might as well recommend fortran, heh or at this point why not just use assembly? since that will help build your "logic"

1. For someone who claims to be working in AI even before that was a buzzword, you don't seem to know what complex logic or even logic is w.r.t AI field as you quote Assembly language for logic programming. I wonder why.
1.1 Anyways Since you so nicely asked, here is a very simple logic deduction exercise for you,
alpha: (P /\ ~P)
beta: P /\ ~ (Q -> P)
gamma: (P V Q <-> ~ P /\ ~Q)
Are all three formulae consistent?
1.1.1 Now that you have an idea of what logic and what logic deduction is, can you find out where such a thing is used in AI? Hint: Deductive Reasoning
1.1.2 Now that you know where logic deduction is used, can you may be think of an application where that is actually used in real world? Hit: Legal and Medical Support
1.2 Now that you know what complex logic is, What is the best language for a beginner to implement Point 1.1 or any logic reasoning?
1.2.1 If you still struggle, here is a hint why I suggest prolog, "best language for a beginner" to "learn logic and flow"
1.3 I hope you wont ask what "flow" is next.

please enlighten me, and please dont mention decades old infra surviving on nothing but prayers and patchwork, or old researchers who are used to it and just wont change, hell, name me just one new (started in the last 10 years) project thats using Lisp

2. Your argument of "please don't mention decades old infra surviving on nothing but prayers and patchwork" and "old researchers who are used to it and just wont change" has no relevance to the point that I made in my comment which is "Learn LISP, Prolog, they are not the cutting edge or anything but when you program in them you will understand logic and flow better than in other languages."
3. Since its established you don't know what logic is,
3.1 It was you who said that prolog, lisp are dead and not used anywhere.
3.2 I said they will help the student understand logic better as they are easy to program.
3.3 Your point of "name me just one new (started in the last 10 years) project thats using Lisp"
3.4 Point 3.1 has no relevance to 3.2 and so 3.3 is your issue not mine.
3.5 But still as you so nicely asked here is one http://www.opusmodus.com, Want more? Here is another https://allegrograph.com

and I would if it wasnt blatantly clear that you are just copy pasting stuff without doing any research, if I wanted that I would stick to shitty medium articles or reddit. I love TE because peeps post what they know instead of just copy pasting rando AI generated crap

4. The only crap that I see here are the one that you have written.
4.1 Just because you cant reply coherently without name-calling or spewing nonsense doesn't mean others cant.
4.2 Example of your nonsense, exhibit below,

Applied AI which is most likely what you are gunning for anyways is easier, but these days it mostly consists of LLM Prompting, maybe some embeddings, some trasfer learning, and thats about it.

4.5 Robotics control is Applied AI, Medical Diagnostics is Applied AI, Legal Support System is Applied AI, none of them are "LLM prompting" or "some embeddings" or "transfer learning", these are full fledged system that encompass a lot of AI technologies.
4.6 Think whatever you want to think, I have nothing more to add.

 

[altair21]

1. For someone who claims to be working in AI even before that was a buzzword, you don't seem to know what complex logic or even logic is w.r.t AI field as you quote Assembly language for logic programming. I wonder why.
1.1 Anyways Since you so nicely asked, here is a very simple logic deduction exercise for you,

let me clarify for OOP as well, I was the same as them, trying to break into Applied AI and research fields before I gave up, Its been years since I have touched any research paper, but I do follow along the developments cursorily. I misunderstood your complex logic part, for that, my bad

alpha: (P /\ ~P)
beta: P /\ ~ (Q -> P)
gamma: (P V Q <-> ~ P /\ ~Q)
Are all three formulae consistent?
1.1.1 Now that you have an idea of what logic and what logic deduction is, can you find out where such a thing is used in AI? Hint: Deductive Reasoning

I had to look up the symbols, but here it goes- first one says P is true and P is not true? so not consistent, second says P is true and not of ( Q is true implies P is true)? which can only happen (assuming Q is true) when P is false, so inconsistent, third one I am way too lazy to attempt. but I dont see anything that says we need lisp/prolog for this.

1.1.2 Now that you know where logic deduction is used, can you may be think of an application where that is actually used in real world? Hit: Legal and Medical Support
1.2 Now that you know what complex logic is, What is the best language for a beginner to implement Point 1.1 or any logic reasoning?
1.2.1 If you still struggle, here is a hint why I suggest prolog, "best language for a beginner" to "learn logic and flow"
1.3 I hope you wont ask what "flow" is next.

yeah you are an idiot, I misunderstood your complex logic part but what exactly something like Prolog/Lisp offers that Python, a way more general and upto date language doesnt offer? and also yeah I dont know about "flow", I mostly went deep into ANNs and their applications, I think in my time, latest breakthrough was Google's BERT for NLP? iirc it was a major breakthrough in encoder models because I remember being excited about it and trying to decipher its paper.

2. Your argument of "please don't mention decades old infra surviving on nothing but prayers and patchwork" and "old researchers who are used to it and just wont change" has no relevance to the point that I made in my comment which is "Learn LISP, Prolog, they are not the cutting edge or anything but when you program in them you will understand logic and flow better than in other languages."

again you havent given one argument about why its better, forget about justifying whether that even makes it worth it learning a general purpose language like Python for a new comer to CS? like bro, they'll probably spend weeks wrapping their head around recursion, using loops effectively, optimising them, what makes you even think that something like Lisp will help a newcomer better than Java/Python where are probably as many tutorials on basic concepts as there are hairs on someone's chest?

3. Since its established you don't know what logic is,
3.1 It was you who said that prolog, lisp are dead and not used anywhere.
3.2 I said they will help the student understand logic better as they are easy to program.
3.3 Your point of "name me just one new (started in the last 10 years) project thats using Lisp"
3.4 Point 3.1 has no relevance to 3.2 and so 3.3 is your issue not mine.

3.5 But still as you so nicely asked here is one http://www.opusmodus.com, Want more? Here is another https://allegrograph.com

I'll give you this I had no idea Lisp was used in these industries, but my point about Lisp being dying still stands, you can look into stack overflow trends, statistics on github, Job openings or any other metric and tell me if I am wrong.

and also, the companies that you listed, may have their initial offering/base products on lisp, but are you really gonna tell me their LLMs etc are built on Lisp?

4. The only crap that I see here are the one that you have written.
4.1 Just because you cant reply coherently without name-calling or spewing nonsense doesn't mean others cant.
4.2 Example of your nonsense, exhibit below,

4.5 Robotics control is Applied AI, Medical Diagnostics is Applied AI, Legal Support System is Applied AI, none of them are "LLM prompting" or "some embeddings" or "transfer learning", these are full fledged system that encompass a lot of AI technologies.

give me one ****ing JD that'll take a Btech grad from a no-name university and expect them to work from day one? if you cant figure out that I was talking from the lens of a new grad, then I dont know what to tell you, I have my juniors, juniors' juniors, my cousins, hell 2024 grads, all looking into trying to breakthrough AI with just a B.Tech, and I havent found any JD from a decent company that does not require either a MS/PHD or some research papers published in good journals to show that you know your shit, and even then they want you to have experience before they'll interview you. And are you really gonna tell me that these fields that you mentioned and their products were not built by MS/PHD folks? specially since you mentioned Robotics Control, are you really gonna tell me that you know organizations that developed their own Reinforcement Learning/NLP/VIsion models on their very ****ing own and whose names are not Boston Dynamics or any other leading robotics company? and are you telling these companies are using Lisp or heaven forbid Prolog instead of building their own custom code on top of something like CPP? and again that they will accept peeps with just a B.S. Degree?

if you really think so, then you are sorely out of touch my guy, all the shit that you just detailed? none of them accept B.S. grads without some stellar research papers/industry experience.

And are you gonna tell that most of these so called new age "AI" startups are not just LLM wrappers with their own custom prompts, and using something like Together/LLama for their embeddings model or what not? and maybe some re-rankers if their usecase is complicated enough

4.6 Think whatever you want to think, I have nothing more to add.

you have added plenty


P.S. do remember to see it from a newcomer's lens and whats practical for them before you respond, most of the stuff that you mentioned? that would be applicable for 2nd or 3rd year students who have gotten their fundamentals strong, know a little bit about memory, OS, DB etc and can make a judgement about where they would like to specialise in

 

[b.life]

Its been years since I have touched any research paper, but I do follow along the developments cursorily.

X.1 What does that have anything to do with what we are talking here?
X.2 Why are you making arbitrary statements?

but I dont see anything that says we need lisp/prolog for this.

1.1 Lets not twist words here, I never said "need", it is you who said it now.
1.2 I gave you an logic deduction exercise to show what logic deduction is as you didn't knew what it was in the first place.
1.3 I also said that writing logic in prolog is easy for beginners.
1.4 I asked you direct question in Point 1.2 of my previous post, for which you have not answered, Why did you not?
1.5 Now again you ask why, even though me stating the obvious.
1.6 If you have problem in understanding what is being said, that is your issue, I don't have to repeat anything.
1.7 To help you even further here is a link to
1.7.1 Stanford course on logic programming, what language they use?(not Proglog but a logic programming language called Epilog) Link
1.7.2 University of Edinburgh course on Logic Programming, what language do they use? Prolog Link
1.7.3 University of Gorningen course on Logic Programming, what language do they use? Prolog Link
1.7.4 Why do these universities use logic programming languages to teach logic programming? How stupid of them?
1.7.5 You said "prolog is dead" but look at all these universities teaching the dead languages and scamming the students. How dare they?

yeah you are an idiot,

X.3 I prove you wrong and you call me names. Great come back. Doorknob is happy that it is no longer the last.
X.4 This makes me wonder
X.4.1 Is throwing insults the only thing he is are capable off?
X.4.2 Does he overcompensate his other deficiencies by throwing insults around?
X.4.3 Is this some kind of coping mechanism for him?

what exactly something like Prolog/Lisp offers that Python,

2.1 Why are you talking about python here,
2.2 Are we or are we not talking about prolog and lisp here?
2.4 Isn't it was C/C++ before this, why change it to python now?
3 Since when did it became a comparison of prolog/lisp with python?

I dont know about "flow", I mostly went deep into ANNs and their applications, I think in my time, latest breakthrough was Google's BERT for NLP? iirc it was a major breakthrough in encoder models because I remember being excited about it and trying to decipher its paper.

X.5 More rambling with no relevance to the topic.
X.6 What does that bring to the discussion?

again you havent given one argument about why its better,

4.1 You willfully ignored what I said.
4.2 Then claim I haven't given any argument.
4.3 If you are going to ignore the argument anyway, why are you asking for one in the first place?
4.4 Why didn't you answer my question?

like bro, they'll probably spend weeks wrapping their head around recursion, using loops effectively, optimising them,

5. Maybe or maybe not, but what is the point here of listing them here?

what makes you even think that something like Lisp will help a newcomer better than Java/Python where are probably as many tutorials on basic concepts as there are hairs on someone's chest?

6.1 Prolog and Lisp also has enough tutorials about them so what is your point here?
6.2 Prolog, Lisp was designed for a specific case, and recommended for the same.
6.3 You are now comparing that specific developed languages with general purpose languages and claiming that general purpose language are better.
6.4 Why are you comparing apples to oranges?

but my point about Lisp being dying still stands, you can look into stack overflow trends, statistics on github, Job openings or any other metric and tell me if I am wrong.

7.1 You are the one who is stating that, so it is on you to prove, why ask me?

and also, the companies that you listed, may have their initial offering/base products on lisp, but are you really gonna tell me their LLMs etc are built on Lisp?

8.1 Lets state some facts here shall we?
8.1.1 You challenged me to list at least one project that is using LISP.
8.1.2 I listed not one but two good projects.
8.1.3 The company website clearly states that they use LISP.
8.1.4 Now you are saying that they didn't build their llm models in lisp.
8.2.1 If you don't believe them then that is on you.
8.2.2 I have nothing to do with your biased views.
8.2.3 Why do you want me to comment on your biased views?
8.3 Now that I have listed two projects you have moved the goal post from listing at least one project to something else.
8.4 Do you even have a goalpost or you make up one as you go along and discard the old ones?
8.5 Why did you move the goalpost now?

give me one ****ing JD that'll take a Btech grad from a no-name university and expect them to work from day one?

9.1 As usual more random statements with no relevance to discussion on hand.

if you cant figure out that I was talking from the lens of a new grad, then I dont know what to tell you, I have my juniors, juniors' juniors, my cousins, hell 2024 grads, all looking into trying to breakthrough AI with just a B.Tech, and I havent found any JD from a decent company that does not require either a MS/PHD or some research papers published in good journals to show that you know your shit, and even then they want you to have experience before they'll interview you.

9.2 Why are you talking about employability of the new graduate?
9.2.1 Where I have given proof that what you initially wrote is nonsense
9.2.2 You made a claim that Applied AI these days is mostly "LLM Prompting, maybe some embeddings, some trasfer learning, and thats about it."
9.2.3 I listed real Applied AI domain examples "Robotics control is Applied AI, Medical Diagnostics is Applied AI, Legal Support System is Applied AI"
9.2.4 How does that translate to employability of graduates? What kind of logic is this?

And are you really gonna tell me that these fields that you mentioned and their products were not built by MS/PHD folks? specially since you mentioned Robotics Control, are you really gonna tell me that you know organizations that developed their own Reinforcement Learning/NLP/VIsion models on their very ****ing own and whose names are not Boston Dynamics or any other leading robotics company? and are you telling these companies are using Lisp or heaven forbid Prolog instead of building their own custom code on top of something like CPP? and again that they will accept peeps with just a B.S. Degree?

if you really think so, then you are sorely out of touch my guy, all the shit that you just detailed? none of them accept B.S. grads without some stellar research papers/industry experience.

9.3 What does listing of real Applied AI domain examples have anything to do with who developed it?
9.3.1 Why all of a sudden you are taking about people who developed it?
9.3.2 Did I claim anything even remotely related to who developed it?
9.4 Where did I tell/talk/imply/hit about who developed robotics control?
9.4.1 What kind of mental gymnastics are you doing here?
9.4.2 Why are you try to link two unrelated things here?
9.5 Where did I talk/imply/hit that these companies are using Lisp or Prolog?
9.5.1 What does listing of real examples of Applied AI have anything to do with prolog and lisp?
9.6 Where did I say that these companies accept graduates with bachelor's degree or even remotely related to employability?
9.7 Are you really incapable of understand what is being said?

And are you gonna tell that most of these so called new age "AI" startups are not just LLM wrappers with their own custom prompts, and using something like Together/LLama for their embeddings model or what not? and maybe some re-rankers if their usecase is complicated enough

10 Why are you talking about "new age AI startups"?
10.1 Where did I talk about them?
10.2 Where did I talk about what they are using or doing?
10.3 Who are these "new age startups" you are talking about and how are they revelavent here?

P.S. do remember to see it from a newcomer's lens and whats practical for them before you respond, most of the stuff that you mentioned? that would be applicable for 2nd or 3rd year students who have gotten their fundamentals strong, know a little bit about memory, OS, DB etc and can make a judgement about where they would like to specialise in

11.1 Why all of a sudden you are talking about newcomers perspective?
11.2 When was there a discussion about newcomers perspective before?
11.3 Why are you shifting the goalpost now?
11.4 Show me in your post where you are talked with newcomers perspective before.

 

[altair21]

11.1 Why all of a sudden you are talking about newcomers perspective?
11.2 When was there a discussion about newcomers perspective before?
11.3 Why are you shifting the goalpost now?
11.4 Show me in your post where you are talked with newcomers perspective before.

know what? I'll reply to your entire essay but before that I'll quote you my posts, though god knows why I need to considering what OOP asked, though if you still cant get the implied context that all of this IS for a newcomer, then join some weekend English classes and develop reading comprehension

Unless you want to be a "glorified" data muncher, there's zero scope in India for any core AI job and this is true for other countries as well, if you really want to get into core AI (which I doubt you want to), A-Levels/SATs/GRE is the way forward, if you can afford it, do well in SATs and build your career in USA, or if money is an issue, do B.Tech here, and then a MS and PHD from America/Europe.

I wont suggest any AI crap degree from even IITs, specially from a no name not even deemed (from a cursory google search) university, all such universities will just scam you out of your money. Get a proper CS degree from a decent established private college like SRM/VIT (yes, these are decent) which have proven track record of placements, or at least have companies visiting their campuses (this honestly should be your #1 benchmark for any college you consider), even if you graduate from IITK/IITD/IISC, you will not be scoring any core AI job unless you consider building mind numbing statistical models AI, you need a MS/PHD for orgs to even consider you, Applied AI which is most likely what you are gunning for anyways is easier, but these days it mostly consists of LLM Prompting, maybe some embeddings, some trasfer learning, and thats about it.

read this all again, very slowly and then read my arguments again real slow then think I don't know maybe all of these arguments are for a newcomer, though one thing I'll give, "applied AI" perhaps wasnt the most correct term I used since it a whole job field, I'll just clarify this here, for me Applied AI in this context was jobs/orgs who will hire a btech grad without any research/industry experience which mostly consists of stuff I detailed above

P.S. do let me know if you still want me to reply to your essay, I will but it will take some time, probably over weekend or next week

 

[ninjenstein]

While I know this is not the answer you asked for, I urge you to let the student to take their own decision based on what they want to do right now.

Many Indian parents force their children to do IIT JEE / Medical as if they're the only two options available, ignoring the child's current interest, thinking the child is not smart enough to decide for themselves, and justifying their own bad decision of forcing the child to do something they don't want to by saying "Careers are flexible anyway, if my child gets interested in something else later they can always change it."

If you actually believe careers are flexible, then let the child choose what they want to do right now, not later. It is a slog and mental hell to do something you don't care about at all for 4 years of your life. Any individual only succeeds if they do what they like. This has been proven time and again through countless scientific studies. Read the book Grit: The Power of Passion and Perseverance to understand the gravity of what I'm talking about. Give it to your child to read also. Since you're on Techenclave I can safely assume you are smarter than the average Indian uncle / aunty. Respect your child's opinions and let them pursue what they want to do right now, and guide them to get into the best college for that.

In case the child is struggling to pinpoint what they're truly passionate about, ask them what they've been consistently doing for a long time. Consistency is a proxy for passion. (It is totally normal to be confused and not know what to do. children are forced to make an irreversible critical important decision at a very young age, irreversible as in irreversible for 4 years, since most colleges don't allow changing your major as you explore your interests and discover new things, unlike universities abroad, which is a folly in the education system of our country.)

I'm not saying you are the typical Indian parent who thinks of their child as an extension of themselves, doesn't believe in their ability, doesn't let them explore, and takes zero effort and interest in understanding their child as a person, but I wanted to put this out here just in case.

 

[altair21]

but coming back to the original topic at hand, in my experience there are 3 types of jobs available for a B.Tech grad in India

1. Smaller Service based shops: They mostly have clients on contract like banks etc, and they mostly do statistical modelling commonly, using decision trees etc for smaller models. They mostly use something like pyspark for data ingestion, some hyper parameter tuning but is mostly based on ML + statistics
2. Big orgs with in-house data teams (I interned in one of these but its been years so my info may be out of data): Here you'll mostly either be a Data Engineer or a Data Scientist or even a cross between them depending on team size/org needs, they'll have their own in-house data, mostly utm stuff, Customer journeys, customer conversions etc, basically you'll again be mostly using ML models to predict customer trends, increase conversions etc, and work on Data Visualisation tools like Tableau and work with internal stakeholders.
3. "New Age Startups": Basically any startup dealing with semantic reasoning, let's say any figma to code converters, any niche-parsers etc: these mostly use LLMs under the hood, use pre-trained models etc, basically while you wont be doing any of the training, you will be using those models for your usecase which also makes for good learning, but its not core AI per-se, like you wont be implementing models from research papers, not training your own models to fit your usecase etc

Pay wise, being a Data Scientist in India is crap, you wont get paid much, but if you put in the time (minimum 4-5 years) your pay will increase exponentially, bigger orgs like ZS might pay more out of the gate but thats not the norm. even Amazon gives a 25L-ish CTC for a data scientist role for new grads but for a SDE role, CTC is much higher.

EDIT: just a clarification on #1, by smaller I meant service based orgs which do not come under WITCH moniker, these can vary from a 50 person shop to a something like ZS, and payscale and complexity of work also varies across the board

 

[b.life]

know what? I'll reply to your entire essay but before that I'll quote you my posts, though god knows why I need to considering what OOP asked, though if you still cant get the implied context that all of this IS for a newcomer, then join some weekend English classes and develop reading comprehension

Since you say that you are always talking with newcomer's perspective, Lets just assume you did so the entire time and also as I don't have much English comprehensive skills can you

1. Tell me where that is in this first reply of yours to mine that perspective, hidden or otherwise?

bruh whats this AI garbage that you are copy pasting? you definitely dont need either LISP or Prolog for AI, and if you really want to learn the extreme basics, then C/C++ is your best bet, Is prolog even being used anywhere? same with Lisp? atleast fact check before you copy paste shit

2. Also in this second reply of yours to mine, where is that perspective?

what exactly do you mean by "complex" logic? like you do know that Lisp is on its deathbed and that prolog is dead? you might as well recommend fortran, heh or at this point why not just use assembly? since that will help build your "logic"

my point above

please enlighten me, and please dont mention decades old infra surviving on nothing but prayers and patchwork, or old researchers who are used to it and just wont change, hell, name me just one new (started in the last 10 years) project thats using Lisp

and I would if it wasnt blatantly clear that you are just copy pasting stuff without doing any research, if I wanted that I would stick to shitty medium articles or reddit. I love TE because peeps post what they know instead of just copy pasting rando AI generated crap

3. All that I read in your previous replies are mockery and name calling. But what do I know. You will show me right?

read this all again, very slowly and then read my arguments again real slow then think I don't know maybe all of these arguments are for a newcomer, though one thing I'll give, "applied AI" perhaps wasnt the most correct term I used since it a whole job field, I'll just clarify this here, for me Applied AI in this context was jobs/orgs who will hire a btech grad without any research/industry experience which mostly consists of stuff I detailed above

4. Your quoted text was posted even before I commented on this thread. And In my first post I agreed on your conclusions about Indian colleges. But where is the newcomer perspective when you were discussing with me?
5. I can only respond to what you said to me, if that has no perspective then what do you expect the response will be?
6. And since you said I am to talk with newcomers perspective, what was the perspective that you understood from my first post in this thread?

P.S. do let me know if you still want me to reply to your essay, I will but it will take some time, probably over weekend or next wee

6. What you do is yours to decide.

 

[altair21]

Since you say that you are always talking with newcomer's perspective, Lets just assume you did so the entire time and also as I don't have much English comprehensive skills can you

1. Tell me where that is in this first reply of yours to mine that perspective, hidden or otherwise?

you are beyond help my guy, if you cant even comprehend that the perspective shifts even in the same convo.

2. Also in this second reply of yours to mine, where is that perspective?

3. All that I read in your previous replies are mockery and name calling. But what do I know. You will show me right?


4. Your quoted text was posted even before I commented on this thread. And In my first post I agreed on your conclusions about Indian colleges. But where is the newcomer perspective when you were discussing with me?
5. I can only respond to what you said to me, if that has no perspective then what do you expect the response will be?
6. And since you said I am to talk with newcomers perspective, what was the perspective that you understood from my first post in this thread?

you literally recommended almost dead languages to a newbie to CS which are used in niche AI applications, which according to the Stackoverflow 2024 developer survey, just occupied a little over 2.3% in the popularity survey to a newcomer on the logic that it will be useful for "complex logic" and "flow"? like why? Like I said before, any newcomer would be busy learning recursion, loops. doing basic stuff like printing patterns etc, and in general just getting their minds thinking in terms of a machine? specially when most of the modern world and the vast majority of AI tutorials/libs etc are built for python and Lisp/Prolog being ancient having just straight out obscure syntax for anyone? like bro come on

 

[b.life]

you are beyond help my guy, if you cant even comprehend that the perspective shifts even in the same convo.

I literally asked you to show but you make a generic statement as usual.

you literally recommended almost dead languages to a newbie to CS which are used in niche AI applications, which according to the Stackoverflow 2024 developer survey, just occupied a little over 2.3% in the popularity survey to a newcomer on the logic that it will be useful for "complex logic" and "flow"? like why? Like I said before, any newcomer would be busy learning recursion, loops. doing basic stuff like printing patterns etc, and in general just getting their minds thinking in terms of a machine? specially when most of the modern world and the vast majority of AI tutorials/libs etc are built for python and Lisp/Prolog being ancient having just straight out obscure syntax for anyone? like bro come on

You keep saying dead language, but refuse any reply that doesn't suit your narrative, so there is no point in this discussion anymore. So good luck.