Breaking Into Tech Without a CS Degree: The Reality Behind Career Transitions in the AI Era

An honest conversation with Prateek Jain about transitioning from B.Com and marketing to data engineering at S&P Global. The real challenges, hiring biases, and whether career switches are still possible in the AI automation era.

July 21, 2025

5 min read

By Rachit Magon

Can you really break into deep tech roles without a computer science degree, or are we just selling false hope to people? While Goldman Sachs reports that AI could automate 300 million jobs, YouTube channels promise six-figure engineering careers after a six-month course.

Today we're cutting through the noise with Prateek Jain, a data engineer at S&P Global who made the jump from B.Com and influencer marketing to architecting scalable data solutions - in just six years. His story isn't a feel-good success narrative. It's an unflinching look at hiring biases, the real impact of AI automation, and whether career transitions are still viable in 2025.

This conversation reveals what actually happens behind the scenes when you try to switch careers without the "right" credentials.

Key Takeaways: The Brutal Truth About Non-CS Career Transitions

The Reality Check:

• Career transitions are possible but significantly harder than YouTube gurus claim
• Hiring bias against non-CS candidates is real, even at companies claiming otherwise
• AI is changing the game but creating new opportunities for those who adapt strategically

Immediate Actions (Next 30 Days):

• Focus on understanding business context, not just coding syntax
• Start building projects that solve real problems, not tutorial clones
• Learn to work alongside AI tools while understanding their limitations

Long-term Strategy (1-2 Years):

• Develop domain expertise that combines technical skills with business understanding
• Position yourself as someone who connects dots between systems, not just builds components
• Build a portfolio that demonstrates problem-solving, not just code generation

Q: Let's start with your journey. You transitioned from B.Com and influencer marketing to data engineering. What triggered that switch?

Prateek: Thanks Rachit for having me. I think it's a very important topic to cover. Yes, I transitioned from B.Com to an engineering role, and I also did a computer science degree during my job - I completed it last year. I've been working in software engineering for six years now, currently as a data engineer.

I started as an influencer marketer back in 2019, and now I'm here in my current role. There was no specific trigger for the switch. Influencer marketing is a good industry and I liked it as well, but it just wasn't for me. I think I wouldn't have been able to do very well there. It was a different sort of fun, but every role isn't for every person.

🔥 ChaiNet's Hot Take: This honest assessment is refreshing. Most career transition stories paint the previous field as terrible to justify the switch. Prateek's recognition that marketing wasn't wrong - just wrong for him - shows the maturity that actually makes career transitions successful.

Q: Goldman Sachs says AI could automate 300 million jobs, but YouTube channels promise engineering careers after six-month courses. What's the reality?

Prateek: That's a very good question. I won't deny that operational and repetitive tasks we were doing before will get automated now, especially with platforms like Replit, Lovable, and others. But we have to understand that even before AI came into the picture, we were automating things.

Organizations and teams who want to do productive work have always automated operational tasks to move on to strategy, thinking, and planning. With AI, it's become simpler for everyone to adopt automation, not just people in software roles who can write code.

Automations have always been there. Yes, if 10 people were needed to automate a task before, maybe now two or three are required. But that doesn't mean the other eight will necessarily lose their jobs. If they adapt to this change, they may get better roles within the same or different organizations because everyone needs talent to drive initiatives.

As for YouTube influencers - learning technology is still a good thing. But if they're promising you'll be able to work the same way as before, that's wrong. If you don't understand Python, React, or these technologies, even if AI writes code, you won't know if it's correct. You can't rely on AI writing 100% correct code.

🔥 ChaiNet's Hot Take: This nuanced view cuts through both the AI doom and the bootcamp hype. The key insight: adaptation beats displacement. Those who learn to work with AI tools while understanding their limitations will survive the transition.

Q: We're seeing claims that AI can write 30% of code being pushed. What are we training junior developers for if AI is writing the code?

Prateek: Every company right now is experimenting with AI code. They're experimenting to see if writing code with AI automatically serves the purpose or not. I think it cannot still do it at 100% capacity because it doesn't have the entire context.

Let's say you have an entire business use case behind a feature, and it may be divided into 10 sources. The code you're working on right now - AI may not be aware of the other 10 sources it's getting data from. It's difficult for AI to have the entire context of everything you're doing in that feature or product to write everything completely on its own.

If it's an algorithmic issue - performance optimization, DSL things - that can easily be done by AI-written code. But if there are things where you need to put in your brain to connect the dots between different application resources, destinations, pipelines, you need human intervention.

Junior developers right now need to be trained on using AI along with getting the job done, not just writing prompts to the editor.

🔥 ChaiNet's Hot Take: This context problem is crucial and often overlooked. AI excels at isolated tasks but struggles with system-wide understanding. This creates a natural moat for engineers who understand how systems connect - exactly the kind of thinking that non-CS backgrounds often bring.

Q: I've noticed web development seems most affected by AI automation, while distributed systems and data engineering seem more protected. Your thoughts?

Prateek: Web development - I'd say yes, it's 80% there with AI writing code. But if you notice, the UI that AI gives is not something you may be able to use for complex applications.

There are two kinds of websites: your landing pages and basic informational sites. AI is doing a great job automating those - it can write good code to get a basic website done, and for most people, that's enough.

But if you want to create something great as a SaaS - good dashboards, panels, or portals - if you ask AI to create something, it will give you very PC-level UI that may not serve your purpose entirely. There will be inefficient code written.

Very basic example: you go to UI coded by Lovable, and every time you open that tab, it will reverify your login. That's a very basic frontend thing, but it still does that.

So it works for some applications, but web development is getting easier with AI, not necessarily better.

🔥 ChaiNet's Hot Take: This distinction between "good enough" and "production-ready" is where many career transitions fail. Understanding the difference - and building skills for the production-ready side - is what separates successful transitions from failed attempts.

Q: I see claims on Reddit about people building tools in 2-3 days with AI and making $10K MRR. What's your take?

Prateek: I think that's really good marketing, but I wouldn't call it a good product. These people who share these claims - I haven't used these products myself, and I haven't seen them being used very openly.

What happens is they ask Lovable to create something, and it can do it within minutes. Then they give it prompts, and in 24 hours they make it presentable. They create hype on Reddit - the posts we see, they're probably doing more posts before that to get users to try it out.

As soon as they release that product, they have some users on board. But that number doesn't do justice to long-term retention. It's maybe MRR for a month, but you don't know what's going on in the backend.

I remember one Twitter post from such a founder who created a product, got funding, and was doing great numbers. Then a few weeks later, he posted that someone brute-forced bad entries into their database, flooding it with irrelevant data. Now he was looking for an actual engineer to help fix it.

Eventually, you'll need engineers.

🔥 ChaiNet's Hot Take: This is the reality check the indie hacker community needs. AI can create impressive demos quickly, but building robust, scalable systems still requires engineering fundamentals. The gap between "working prototype" and "production system" remains huge.

Q: Let's talk about hiring. How did you overcome the initial bias against non-CS candidates?

Prateek: The first couple of switches were not easy, and this was a very common question from every hiring manager: "Why are you trying to move from B.Com to this role? You haven't done a BTech."

I never tried to run away from it or hide it. I was honest with whatever recruiters asked me, and I think some recruiters or companies appreciated that I was trying to self-learn and move because that gives me more motivation to succeed in the industry.

Being from a commerce background gives me a little edge because I try to understand the business behind it instead of just looking at the code or thinking about how to write code.

🔥 ChaiNet's Hot Take: This business perspective is actually becoming more valuable in the AI era. As AI handles more implementation, the ability to understand business requirements and translate them into technical solutions becomes the differentiator.

Q: There's a debate about skills versus degrees. Do you unconsciously favor CS graduates when hiring?

Prateek: I don't think I do it because education is the last segment I see in a resume. I first check experience and projects.

But I'll be honest - I was working at a startup 3-4 years back where they started hiring people from IITs, even freshers, but not for senior roles. When I asked the CEO and CTO why there was an IIT check, after asking multiple times, they replied: "Because we need funding, so we need to show that we have IITs working for us, even if it's freshers."

That's still the reality for many startups. Companies claim they don't discriminate, but they would prefer someone with a CS degree.

I remember one case where a candidate asked me during the interview: "If I get rejected, will it be because of my non-CS degree? I'm from a BBA background." I told him I didn't even know his educational background until then - education would be the last thing I'd judge him on.

🔥 ChaiNet's Hot Take: This brutal honesty about startup hiring practices is valuable. The funding angle explains why many companies publicly claim "no degree requirements" while privately filtering for prestigious credentials. It's not about technical capability - it's about investor perception.

Q: Are we automating ourselves out of our future? Your frameworks and lead management systems are exactly what AI coding assistants are now doing.

Prateek: There are two angles to it. If you're trying to build what's already out there, then AI can easily do it. But if you're trying to build something very unique with a specific feature or functionality that wasn't there before, I doubt AI can build it from scratch.

Like the Lego comment earlier - if you're building an entirely new product, some Lego pieces common among similar platforms can be built by AI, but AI can't build a new invention out of nowhere. If it hasn't done it successfully before, I don't think it can properly build that new feature. It will have bugs, and you'll need an engineer to build that innovation from scratch.

AI has automated a lot of everyday tasks, yes. But if you're trying to do something new, AI won't help you there much.

🔥 ChaiNet's Hot Take: This innovation gap is where career changers can find their niche. Coming from different backgrounds, they often see problems differently and can identify unique solutions that AI, trained on existing patterns, might miss.

Q: CEOs like Jensen Huang say no one will write code eventually, while Marc Benioff says Salesforce won't hire more engineers. Are they creating hype?

Prateek: These people are pretty high in the chain - I can't entirely deny what they say. But when they say people won't write code anymore, they don't mean software engineering will go away.

AI will reduce the amount and type of code we write and save us time. It's giving us time to plan, strategize, and actually design systems. If we design the system and AI writes code for us, we can ship products way faster.

But people getting afraid that "if we can't write code, we don't have a job" - that's the wrong tangent. Software engineering will still exist as a field, but the meaning of writing code as we know it might change.

🔥 ChaiNet's Hot Take: This reframing is crucial for career transitions. The focus shifts from "learning to code" to "learning to architect and design." This higher-level thinking is actually more accessible to people from business backgrounds who already think in terms of processes and systems.

Q: Will we see a golden age of solo builders, or just a tsunami of mediocre AI-generated apps?

Prateek: Because of AI and tools like Cursor and Windsurf, people will definitely try to build something of their own. Everyone is trying to achieve certain goals, and most of those goals are similar.

If the product they're building doesn't solve a certain purpose, I don't think it will do well. Even if you use AI and build something, if it's not solving a good problem, it won't go a long way. You can create all the hype you want and do personal branding, but if the product isn't helping achieve a goal, it's not a good product.

Even if you tell people "I built this in 48 hours" and get initial trial users with some MRR, if users don't stay on your platform after one month, it's not worth it.

I still think solopreneurship is subjective when it comes to software products. Even if you build everything with AI initially, there will come a point where you need to optimize processes, and that's where you need real engineers. You can't keep using everything through Lovable - you have to move out.

For an MVP it might be okay, but eventually you'll need a team, real builders. You cannot have founders just roaming around.

🔥 ChaiNet's Hot Take: This insight about the MVP-to-scale transition is where many AI-generated businesses will fail. The engineers who understand this transition - particularly those with business backgrounds who can see the operational challenges - will become invaluable.

Q: Five years from now, with AI agents handling data pipelines and advanced low-code tools, how do you plan to stay relevant?

Prateek: One thing I've realized about data engineering is that it's never about the tools we use. It's always about the ability to understand the data and how it connects with your actual application.

Data engineering isn't always for post-analytics - you can also use data engineering for incoming streams to your application. Even if AI automates 50% of tasks, you still need planners and designers to get work done in a certain way. You can't always rely on AI to make decisions for you.

Data engineering itself - tools are just a means to get it done. Data engineering is a lot more than just tools. We need custom logic and requirements. Tools are just one very small part of the game.

🔥 ChaiNet's Hot Take: This systems thinking - understanding how data flows through organizations and connects business needs with technical implementation - is exactly what makes career changers valuable. It's not about the tools; it's about understanding the bigger picture.

The Uncomfortable Truth About Career Transitions

Prateek's journey reveals several uncomfortable truths about breaking into tech without a CS degree:

The Good News:

• It's still possible, but requires genuine effort and strategic thinking
• Business backgrounds provide valuable perspective that's becoming more important
• AI creates new opportunities for those who understand its limitations

The Bad News:

• Hiring bias is real and often tied to funding/investor perceptions
• The "six-month bootcamp to six-figure salary" promise is largely false
• Success requires understanding systems, not just writing code

The Strategy:

• Focus on connecting business problems with technical solutions
• Develop deep understanding of how systems work, not just how to use tools
• Position yourself as someone who can guide AI tools, not compete with them

Prateek's path from B.Com and marketing to data engineering at S&P Global wasn't about following a bootcamp curriculum. It was about developing business intuition, technical depth, and the ability to see how systems connect - skills that become more valuable as AI handles routine implementation.

The question isn't whether career transitions are possible in the AI era. It's whether you're willing to do the real work of understanding systems deeply enough to guide them, rather than just implement them.

Connect with Prateek: You can find him on LinkedIn for career guidance and insights into data engineering career paths.

Career transitions in tech remain possible, but the bar is rising. Success goes to those who combine business understanding with technical depth - not those chasing quick wins with AI-generated code.