Give yourself 6–8 weeks of structured preparation. Start with fundamentals — scalability basics, SQL vs NoSQL, caching, load balancing, message queues and API design. Then study 12–15 classic designs such as URL shorteners, chat apps and rate limiters, practising by drawing each architecture while explaining your trade-offs aloud. Finish with 2–3 mock interviews on a 35–40 minute timer, because most Indian product companies expect you to design and defend a system in real time.

Use a repeatable framework: (1) clarify functional and non-functional requirements, (2) estimate scale — users, QPS, storage, (3) define the API, (4) sketch the high-level design, (5) deep-dive the data model, database choice, caching and queues, and (6) address bottlenecks with sharding, replication or load balancing while stating trade-offs. Interviewers care less about a "perfect" answer and more about how structured and trade-off aware your thinking is under time pressure.

The classics repeat across companies: design a URL shortener, a WhatsApp-style chat app, an Instagram-style news feed, a rate limiter, a notification system, a BookMyShow-style ticket booking platform, a video streaming service, a Dropbox-style file store, and a web crawler. Alongside these, expect theory follow-ups on caching strategies, SQL vs NoSQL, CAP theorem, consistency vs availability, and load balancing. Indian startups and product companies often frame these around their own domain — food delivery, payments, edtech.

The most widely recommended system design interview book is System Design Interview by Alex Xu. Volume 1 covers the core building blocks — scaling, caching, load balancing, databases — through step-by-step designs like a URL shortener and rate limiter, while Volume 2 goes deeper into advanced topics such as consistent hashing, proxies and notification systems. Use it for structure, then reinforce it by designing unfamiliar systems yourself and talking through trade-offs, since a book alone rarely gets you over the line.

Grokking the System Design Interview is a popular course that teaches reusable patterns through around 15 classic design problems such as URL shorteners, chat systems and news feeds, making it a gentle starting point for beginners. It is useful for building a template-driven approach, but on its own it is not enough — interviewers deliberately pick unfamiliar systems to test adaptability. Combine it with a good book, regular self-practice on a whiteboard or design tool, and at least a couple of mock interviews with feedback.

Work pattern-wise over 8–10 weeks: master arrays, strings, hashing, linked lists, stacks, queues, trees, graphs and heaps, then the key techniques — binary search, two pointers, sliding window, recursion, backtracking and core dynamic programming. Solve roughly 150–200 well-chosen problems instead of hundreds at random, maintain a mistake log, and redo failed problems a week later. Close with timed mock interviews where you think out loud, since communication carries almost as much weight as the optimal solution.

Freshers are usually tested on arrays and strings, hashing, two-pointer problems, linked list operations, stacks and queues, binary search, sorting, recursion, basic tree and graph traversals, plus time and space complexity. Service companies such as TCS, Infosys and Accenture mostly screen fundamentals and patterns, while product companies like Amazon, Microsoft and Flipkart go deeper into medium-level problems and occasionally light DP. Practise explaining your approach before coding — freshers are judged heavily on clarity of thought.

LeetCode's company-tagged problems, GeeksforGeeks interview experience archives and company-wise question lists are the most reliable starting points. If you are short on time, a curated interview prep pack that compiles frequently asked DSA interview questions and answers with complexity analysis saves weeks of scattered searching. Whatever source you use, do not memorise — understand the pattern behind each solution, re-solve problems from scratch after a few days, and practise writing working code without autocomplete.

Use whichever language you can code in fastest with the fewest bugs — interviewers evaluate logic and complexity, not the language. That said, DSA interview questions in Java are a very common combination in India because enterprise and product backends run heavily on Java, and its Collections framework (HashMap, PriorityQueue, Deque) makes implementing standard structures quick. Python is more concise, C++ is fastest for tight constraints. Pick one, learn its standard library well, and practise speaking your approach while coding.

Prompt engineering is the practice of structuring instructions, context, examples and constraints so AI models like GPT or Claude produce accurate, consistent, usable outputs. It covers techniques such as zero-shot and few-shot prompting, chain-of-thought reasoning, role-based prompts, and grounding responses in retrieved data (RAG) to reduce hallucinations. For software engineers and data professionals in India, it has become a core GenAI interview skill — hiring teams increasingly test whether you can make LLMs reliable inside real products, not just chat with them.

Follow a simple structure: assign a role ("act as a senior code reviewer"), provide full context and constraints, state the exact task and output format, and include one or two examples of what good output looks like. For complex problems, ask the model to reason step-by-step, then iterate — critique the response and refine the prompt instead of accepting the first attempt. Save your best-performing prompts as reusable templates; consistent iteration is what separates real prompt engineering from casual chatting.

Judge any prompt engineering course on four things: hands-on projects rather than pure theory, coverage of core techniques — few-shot prompting, chain-of-thought, RAG and hallucination handling — content updated for current models, and practical use cases like coding, content and data workflows. Free intro courses are fine to test your interest, but if you are targeting GenAI roles, choose a structured bundle with projects and interview-focused material, since Indian hiring teams increasingly assess prompting through live tasks rather than certificates.

A free prompt engineering course with certificate is a good starting point, not a finishing line. It helps you build fundamentals quickly and adds a resume signal, but recruiters for AI/GenAI roles shortlist on demonstrated skill — a GitHub project such as a RAG chatbot, prompt templates with measurable improvements, or LLM-powered features built into real applications. Use the certificate course to learn the basics, then invest in a project portfolio and mock interviews; that combination converts far better than a certificate alone.

Yes, demand is growing fast as Indian startups, global capability centres and IT services firms build GenAI products — but standalone "Prompt Engineer" titles are still rare. Most openings are for AI/GenAI engineers, software engineers and data engineers who can prompt well and also build on LLM APIs, RAG pipelines and evaluation workflows. The practical route is upskilling from an existing base — backend, full-stack or data — into GenAI, which is exactly the switch many mid-level engineers are making in the current hiring cycle.

Salaries vary widely by company, city and skill depth. Standalone prompt engineering roles in India typically start around ₹6–12 LPA for candidates with strong fundamentals, while experienced AI/GenAI engineers who combine prompt engineering with Python, RAG and LLM integration skills commonly earn ₹20–40 LPA at product companies and global capability centres. Since dedicated prompt-only titles are limited, the highest earnings usually go to engineers who pair prompting skills with solid software engineering fundamentals — so build both together.