Career guidance, without the LinkedIn advice.
I run engineering at a fintech and have spent a decade hiring, promoting and occasionally failing to keep good engineers. Most of the questions I get have an honest answer that takes two minutes. Those are below, free. If yours is not one of them, book a session.
Start with the common questions
These are the five I am asked most. Read the answer before you book anything — you may not need to.
Which programming language should I learn in 2026?
It depends on what you want within twelve months, not on which language is best.
- Most job openings in India — Java. The enterprise market is enormous and not going anywhere.
- You want to work in AI — Python. Not optional, not negotiable.
- Highest salary, least competition — Go. Strong demand, fewer people who know it well.
- Build a product alone — Node. One language across frontend and backend.
The honest part: the language matters far less than picking one and going deep. Engineers who know one language properly get hired over engineers who know four badly.
Do I need an ML degree to become an AI engineer?
No. This is the most common wrong belief I encounter.
Most production AI work is backend engineering: calling model APIs, building retrieval that returns the right documents, handling the model being wrong, evaluating whether a change helped. If you can build and operate a reliable service, you are most of the way there.
Training models from scratch is a real job, but it is a much smaller market and it does want the degree. Almost nobody hiring "AI engineers" means that.
Is it too late to switch to AI engineering?
No, and your backend experience is an advantage rather than something to apologise for.
The scarce skill is not prompting — it is knowing how to make an unreliable dependency safe in production. Retries, timeouts, fallbacks, observability, blast radius. Backend engineers already do this; the model is just a new kind of flaky dependency.
What you need to add is narrow: how retrieval works, how to evaluate output, and where a human has to stay in the loop.
How do I get from senior engineer to lead?
Stop being the person who writes the most code and become the person whose absence is felt.
Concretely, that is three things: own an outcome rather than a ticket, make the engineers around you faster, and be trusted to make a call without escalating it. The promotion usually follows the behaviour by six to twelve months, not the other way around.
Most people who stall at senior are still optimising for individual output, which is exactly what got them to senior and is exactly what stops working next.
Startup or big company?
Early on, optimise for the quality of the engineers around you. Not the logo, not the equity, not the title.
You learn the craft from people better than you, and that is mostly a function of who is on your team rather than the company's size. A well-run startup and a good team at a large company both work. A badly-run version of either costs you two years and you will not notice until you interview elsewhere.
The equity question: treat startup equity as worth zero when deciding. If it pays out, excellent. Do not take a salary cut for a lottery ticket you cannot price.
Still stuck? Find where you are
Pick the one that sounds most like you. I will point you at what to read first, and tell you whether a session is worth your money.
When a session is actually worth it
I would rather you read the answers above and not pay me. A session earns its money when the question is specific to your situation and a generic answer will not do:
- You have two concrete offers and need help reading what each one actually does to your next five years
- You have been at the same level for three years and cannot work out what is blocking you
- You are choosing between going deeper technically and moving into management, and both look plausible
- You want someone to review your CV and tell you why it is not converting — I read these for a living
- You are preparing for system design interviews and want a real one, not a checklist
Not worth booking for: which language to learn, whether to do a course, or whether AI will take your job. Those are answered above and I will just repeat them.
Free things worth your time
Which language should you learn? — two-minute quiz
Five questions, a recommendation based on your goal rather than on what is trending.
Java vs Python vs Go vs Node
Job volume, salary ceiling and AI relevance for each, with the honest verdict.
How to become an AI engineer
RAG, vector databases, agents and MCP — what matters and what you can skip.
Everything else I have written
System design, architecture and AI engineering, mostly written from things that broke.