Building a career with AI skills in India: what to learn and a 90-day plan

Ask ten people how to build a career in AI and you'll get ten answers, most of them involving a course and a salary figure. This post does neither. It lays out the realistic paths into AI work in India, the skills each one needs, how to prove those skills, and a 90-day plan you can start today.

A bit of context on me: I build AI systems for a living. Our AI video studio was built by one developer working with AI coding agents, and the channel it powers got 1,22,541 views in 7 days (Oct 11–17, 2026). Everything below is what I'd tell a student or colleague asking where to start.

The four AI career paths

You can move between them. Many good AI leaders started as builders, and many creators pick up builder skills along the way.

Skills every path needs: AI fluency

• Prompting with intent: give context, examples and a format; ask for structured output.
• Verification: AI is confidently wrong sometimes. Check facts, test code, read sources.
• Privacy and ethics: know what you must never paste into a tool, and respect consent and copyright.
• Workflow thinking: break a task into steps, decide which steps AI does and which you do.

That last one is underrated. Our studio isn't one magic model. It's 11 stages, each with checks, and a person approving the result. Employers want people who think like that.

Path 1: builder

The most in-demand builder skills today are about using models well in real software, not training them from scratch:

1. Python fundamentals, Git and the command line.
2. Calling AI APIs securely and cheaply (keys in environment variables, caching, retries).
3. Retrieval-augmented generation: embeddings, vector databases, citations, "not found" handling.
4. Agents: function calling, tool use, frameworks like LangGraph or Google's ADK.
5. Evaluation: test sets and metrics so you know if a change made things better.
6. Deployment: an API (for example FastAPI), a simple front end, a free hosting tier.
7. Local models: running open models with Ollama is a useful differentiator. See local AI vs paid APIs.

Path 2: AI-enabled professional

You don't need to change careers to benefit from AI. You need to be the person on your team who uses it best:

• Automate the boring 20% of your week: reports, summaries, first drafts, data cleaning.
• Build small tools without code: custom assistants (such as Gems), document Q&A over your team's files, spreadsheet automations.
• Measure it: hours saved per month is a concrete line for your appraisal or resume.

Path 3: AI leader

If you manage people, budgets or clients, the valuable skill is judgement: which problems suit AI, which solution fits, what it costs, what can go wrong. Google's Generative AI Leader certification is built for this: business-level, no coding, covering gen AI fundamentals, Google Cloud's offerings, techniques like grounding and RAG, and business strategy including secure and responsible AI. Read the full exam guide and how to prepare.

Path 4: AI creator

AI has made it possible for one person to run a channel that used to need a team. It's a real skill set: topic research, scripting, a production pipeline, quality checks, packaging, and understanding YouTube's rules. Start with how to make AI YouTube videos and can AI videos be monetized. Creator income is never guaranteed; treat it as a skill and portfolio first.

Proof: the portfolio that gets interviews

Certificates show you studied. A portfolio shows you can do the work. Aim for:

• 3 deployed projects with public links, each solving a real problem;
• clean GitHub repos with a README that explains the problem, the approach, how to run it and its limitations;
• a 2-minute demo video per project;
• a short write-up of one hard bug or design decision;
• an honest resume: what you built, with what, and the measurable result.

A 90-day plan

Non-builders can swap days 16–70 for three work case studies: a process you automated, a no-code assistant you built for your team, and a business case for an AI pilot.

How to talk about AI skills in an interview

Interviewers have heard "I know ChatGPT" many times. Stand out by being specific:

• Problem, approach, result. "Our team spent hours on weekly reports. I built a document Q&A assistant with citations; it now drafts the report and we review it."
• Trade-offs. Explain why you chose a cloud model over a local one (or the reverse), and what it cost.
• Failure handling. Describe how your app behaves when the model is wrong or the answer isn't in the documents. This impresses more than a perfect demo.
• Evaluation. Say how you measured quality: a test set, a checklist, user feedback.
• Responsible use. Mention privacy, consent and what you deliberately didn't automate.

What to ignore

• Salary screenshots and "guaranteed placement" ads. No honest course can guarantee a job.
• Tool-of-the-week FOMO. Fundamentals outlast tools.
• Certificate collecting without projects.

Where IndMinds fits

• College students and freshers: Practical Gen AI for College Students: 12 weeks, 3 deployed apps, resume, GitHub and interview practice.
• Managers and IT professionals: Generative AI Leader certification prep: independent, live, with labs and two mock exams.
• Creators: the ₹999 weekend workshop, where you build one AI video end to end.

Pick one path, follow the 90 days, and let your projects speak.