Hey everyone,
I’m looking for some honest advice and guidance.
I’m currently pursuing BTech in CSE and I’m in my 3rd year. But honestly, I feel quite stuck. I want to become a Software Engineer, but I’m struggling with a big question:
Is SDE still a good career choice in the age of AI?
AI is improving incredibly fast, and we’re constantly hearing about AI replacing junior developers and reducing the need for entry-level engineers. As someone who hasn’t built strong technical skills yet, this honestly makes me wonder whether I’m already too late.
I know that choosing SDE can still be a great option, but I’m confused about what I should actually focus on in 2026.
Should I focus heavily on DSA and interview preparation?
Should I learn full-stack development?
Should I focus on AI/ML instead?
Or should I learn how to use AI effectively as a developer?
If you’re already working as a software engineer, are a recent graduate, or have gone through a similar phase, I’d genuinely appreciate your perspective.
What would you recommend to a 3rd-year CSE student starting almost from scratch today?
I’m not looking for motivation. I’m looking for practical advice about what to learn, what to avoid, and how to realistically prepare for getting my first software engineering job.
Thanks in advance! 🙌
Top comments (2)
Practical answers to your four questions, in order:
DSA: yes, but treat it as a gate, not the skill. For campus placements you can't skip it, so do a focused pass on patterns rather than 500 random problems — then stop. Nobody's day job is inverting binary trees, and interviews increasingly probe whether you can reason through a system, not whether you memoized a DP table.
Full-stack: this is the one that compounds. AI has made writing medium-quality code for well-defined problems cheap. What it hasn't made cheap is owning something end to end: deciding what to build, wiring auth and deployment, debugging why it works locally and dies in production. Pick a real problem and ship it — real users or at least real data, deployed on a VPS, a domain, some monitoring. Two projects like that beat ten tutorial clones, because the clones collapse the moment an interviewer asks "why Postgres here?"
AI/ML as a specialization: not yet. Entry ML roles want serious math and are more gatekept than SWE, not less. What you should learn is using AI tools well — as a multiplier on the engineering above, not a substitute. I run AI agents that write code all day, and the people worth hiring are the ones who can review what the agent wrote, spot the wrong assumption, and fix the part the agent can't see.
One honest warning: the junior market is harder than it was in 2021, so internships matter more than ever. If you can get one — any company, as long as the work is real — the gap between "student who built projects" and "student who maintained something with users and deadlines" is enormous in interviews. You have a year: shipped project + focused DSA prep + one internship is a realistic combo.
Thankyou so much for the guidance 🙌 The idea of focusing on DSA for the placement gate, building real end-to-end projects, learning to use AI as a tool, and trying to get an internship makes a lot more sense to me.
I have roughly a year left, so if you were in my position and starting almost from scratch today, what tech stack and roadmap would you personally follow to become internship-ready?