How to become a data scientist in India: routes after Class 10, Class 12 and graduation

For students in Classes 9 to 12, graduates thinking of a switch, and their parents. Rules read in official sources on 5 November 2026. Admission rules change every cycle, so check the current notice before you apply.

How do you become a data scientist in India?

Quick answer

Most data scientists in India get there through a degree heavy in maths, statistics or computing, then projects, internships and, often, a master's. Engineering is one route among several. The subject that matters most is Mathematics.

• Science with Physics and Maths keeps the widest set of doors open, including B.Tech through JEE Main (under JoSAA's 2026 rules).[4]
• Commerce or humanities with Mathematics or Applied Mathematics can still lead to statistics and data science degrees, such as the Indian Statistical Institute's (ISI) Bachelor of Statistical Data Science. That is ISI's 2026-27 rule, which asks for Class 12 passed in 2025 or 2026; ISI has not published 2027 rules that we could find.[6]
• No Maths in Class 12? The IIT Madras BS in Data Science is open to any stream, though it expects Maths and English in Class 10, and you will have to work hard on the maths.[5]

Your first job may well be titled "data analyst". That is normal. Read on for the rule that fits your starting point.

At a glance

What does a data scientist actually do?

A data scientist turns a vague question ("Why are customers leaving?" "How many beds will this hospital need next winter?") into something data can answer, then answers it honestly. O*NET, the US government's occupation database, lists the tasks of a data scientist in plain terms.[1] Grouped into stages:

• Framing the problem. Talking to the people who own a business or policy problem, and working out which questions data can actually help with.
• Getting and cleaning data. Collecting data, sometimes designing surveys or choosing samples, then cleaning and reshaping it. This is often the longest part of the job.
• Analysing and modelling. Looking for relationships and trends, building statistical or machine-learning models, and testing and comparing them with proper performance measures.
• Communicating. Making charts, writing reports, presenting to managers and recommending what to do next.
• Keeping up. Reading research and writing new code as methods change.

India has its own national qualification for the role, listed on the National Qualifications Register. Its units, in the page's own titles, are: "Collect & process data and ensure data quality standards are achieved"; "Perform exploratory and Inferential analysis of the data"; "Perform research, design & validation of algorithmic models"; "Risk Analysis & Mitigation for Algorithmic Models"; and "Define business outcomes and create visualizations from results of the analysis". The register page is labelled as a beta version.[3] The risk unit stands out: a model that gets lending or hiring decisions wrong can harm real people, and India's framework counts checking for that as part of the job.

A representative working day (illustrative composite)

This is not a real person's diary. It is a composite drawn from the task lists above, to show the rhythm of the work.

• 9:30 Stand-up with the product team. Yesterday's model flagged too many genuine customers as "likely to leave". You agree to investigate.
• 10:00 to 12:30 Write SQL to pull six months of usage data. One city's records are duplicated, so you fix that and document what you did.
• 14:00 to 16:00 Retrain the model on the cleaned data, compare it with last week's version, and check whether it does worse for any group of customers.
• 16:30 Make two simple charts and write a short note: what changed, how confident you are, and what you would not yet act on.
• 17:30 Thirty minutes reading about a new evaluation method a colleague shared.

Most of that day is careful checking and clear writing, with little "AI magic" in it.

Data scientist vs data analyst vs ML engineer vs data engineer

These titles overlap, and companies use them loosely. As a rough guide:

This table is Edukrish's plain-language summary of how the roles usually differ. It is a rough guide, not an official definition, and job ads vary.

Skills and aptitudes

Technical

• Statistics and probability. This is the core. You need to know when a pattern is real and when it is noise.
• Mathematics. Linear algebra and calculus become important for machine learning.
• Programming. Python and SQL are the most common starting points. O*NET's software list for the job also names tools such as SAS and TensorFlow.[1]
• Data handling. Cleaning, joining and checking data, which is unglamorous and essential.

Non-technical. O*NET's top-listed skills for the job include critical thinking, reading comprehension, active learning and speaking.[1] In practice that means asking good questions, reading carefully, learning new tools on your own, and explaining a result to someone who doesn't care about the method.

How to test your interest early (before you choose a stream or a degree)

1. Ask a question of real data. Take a public dataset (your city's rainfall, cricket scores, exam results published by your board) and try to answer one question with a spreadsheet. Did you enjoy the checking as much as the answer?
2. Try ten hours of beginner Python. Free courses are widely available. If debugging feels satisfying rather than miserable, that is a good sign.
3. Notice how you feel about statistics in school maths. Data science is mostly applied statistics. If probability chapters bore you, look at the Data Analyst or Business Analyst profiles too.
4. Take Edukrish's free interest assessment to see whether your interests line up with Investigative and Conventional work.

Routes in: what you need from where you are now

About the IIT Madras BS. It is an IIT degree taught online with in-person exams. Most people enter through a qualifier rather than a JEE rank: four weeks of online coursework and weekly graded assignments in four foundation courses (English I, Mathematics for Data Science I, Statistics for Data Science I, and Computational Thinking), followed by an in-person qualifier exam. To sit the exam you need a minimum average assignment score in each course, which is 40% for general learners.[5] A separate JEE entry exists: candidates eligible to appear for the most recent JEE Advanced are admitted directly to the foundation level. That is JEE Advanced eligibility, not a JEE Main score.[5] Applications for the January 2027 term were open when we read the page, with 29 January 2027 shown as the last date.[5] Because it is flexible and open to any stream, many students take it alongside another degree. It still demands steady, self-driven work.

Step-by-step roadmap: Class 9 to first job

1. Class 9 and 10. Take Maths seriously. If your board offers Standard and Basic Maths in Class 10, ask your school which one you need in order to take Mathematics in Class 11. CBSE's 2020 circular says students who passed Basic Mathematics in Class 10 are not eligible for Mathematics (041) in Classes 11 and 12; CBSE has relaxed this in some years, and a news report indicates the relaxation continued for 2025-26,[13] but the 2026-27 position is not confirmed here, so ask your school.[11] Try small data projects for fun.
2. End of Class 10. Choose a stream with Mathematics. If you choose Commerce, ask whether your school offers Mathematics or Applied Mathematics, and know that some programmes treat them differently.[6][7] In CBSE, a request to change Class 12 subjects must be made before 15 July of the Class 12 session.[12]
3. Class 11 and 12. Build basic Python and spreadsheet skills. Keep two or three exams on your calendar (JEE Main, CUET-UG, the ISI admission test) and note their application windows. You can apply for the IIT Madras qualifier once you have appeared for your Class 11 final exams.[5] For the 2027 cycle, JEE Main Session 1 applications close on 22 November 2026 (a late window runs 23 to 26 November, with an extra ₹500),[4] and CUET (UG) 2027 had not been announced when we checked on 5 November.[7]
4. Undergraduate years. Pick a degree with real statistics and maths depth (see the next section). From the second year, do one solid project a semester and aim for at least one internship. Keep your code and write-ups in a public portfolio.
5. After the degree. Many graduates start as data analysts or junior data scientists. Others go straight into a master's: MSc Statistics or Data Science, M.Tech, or an analytics MBA. In the US, O*NET's survey splits almost evenly between bachelor's and master's as the expected level,[1] and BLS says data scientists typically need at least a bachelor's, with some employers wanting a master's or doctorate.[2] Expect Indian employers to vary in the same way.
6. First job and beyond. Data science changes fast. Plan to keep learning throughout your career.

Choosing a course: what to check

Many programmes now carry "data science" or "AI" in their name. The label tells you little. Check these:

• Maths and statistics depth. Look at the syllabus. Are there full courses in probability, statistical inference, linear algebra and optimisation, or only "introduction to analytics"?
• Programming and projects. Are there graded projects on real datasets, not just quizzes?
• Internships and placements. Ask for the last two years' placement reports by programme, not just the institution's highest package.
• Faculty. Do the teachers publish or practise in statistics, CS or data science?
• Recognition.
• Check that the university is not on UGC's list of fake universities (see the Edukrish copy and the live UGC link).[9]
• For B.Tech and other technical programmes at affiliated or private colleges, ask for the institute's current AICTE Extension of Approval letter.
• For online degrees, check that the awarding university is itself recognised to award degrees.
• Cost against outcome. Compare total fees with published placement data. Edukrish's colleges pages are a place to start, along with each college's own published placement figures.

Realities and challenges

• Entry-level competition is real. Many people chase the title "data scientist", so first jobs are often analyst roles. A portfolio of finished, well-explained projects separates candidates more than certificates do. (This is Edukrish's editorial judgement, not a measured statistic.)
• Demand is genuine, but often overstated. The evidence, and its limits:
• Employer survey. In the World Economic Forum's Future of Jobs Report 2025, the India profile names Big Data Specialists and AI and Machine Learning Specialists among the roles employers expect to grow fastest.[10] That is a survey of what employers expect, not a count of jobs.
• US projections. BLS projects employment of data scientists in the US to grow 35% from 2025 to 2035, with about 24,800 openings a year on average.[2] That is US data. We have not found a comparable official projection for India.
• Treat dramatic figures with caution. Be wary of large opening figures that don't link to a primary source.
• Continual learning. Tools change every few years. The statistics underneath changes slowly, which is another reason to get the fundamentals right.
• Ethics and responsibility. Models used for credit, hiring or health can be unfair. India's national qualification includes risk analysis of models as a core unit.[3]

Pay: how to read salary claims

You will see confident salary figures for data scientists all over the internet. Edukrish does not give one here. We have not found any official Indian survey that publishes pay for "data scientist" as an occupation. A likely reason is that India's National Classification of Occupations (2015) does not appear to have a code for the role, and the national qualification record we read gives no occupation code either.[3] Surveys that sort workers by occupation code can't easily pick data scientists out. (That is our reasoning, not a published statement.)

What you can use responsibly:

• Placement data from colleges. Median salary of placed graduates is a starting-salary signal for a course, not a promise for the job. (The pay figures on our Data Scientist profile are job-site estimates, so read them with the cautions in the next bullet.)
• Job-site figures (from sites such as Indeed, Glassdoor and AmbitionBox) are self-reported or based on job ads. Check the sample size, the experience level, the city and the date. If two figures on the same page disagree, trust neither.
• US figures are not Indian figures. The US median for data scientists was $120,230 in May 2025.[2] Don't convert it to rupees and expect it in Pune.

Related careers

• Data Analyst: the most common entry point, open to commerce and science students.
• AI/ML Engineer: more software engineering, less statistics.
• Actuary: statistics applied to risk, with its own exam route.
• Economist: data and models applied to policy and markets.
• Business Analyst: closer to business decisions and processes.
• Software Engineer: if you enjoy building more than analysing.

Next steps with Edukrish

• Read the full Data Scientist career profile: pay estimates, entry routes and outlook.
• Take the free career interest assessment: sign up, then open it from your dashboard to see how well data science fits your interests and values, with reasons. It is 38 statements and takes about six minutes.
• Compare careers: save Data Scientist, Data Analyst and AI/ML Engineer, then compare them side by side in your shortlist.
• Check exam pages for dates and rules: JEE Main, CUET-UG, or browse all exams.
• Explore colleges and what they report: /colleges.
• Open My practical pathways (/app/pathways) to see which routes your current subjects keep open.

How we checked

Sources were read on 5 November 2026 unless stated. Eligibility rules come from official bulletins and institute pages. US figures are labelled as US data. Edukrish did not use any job-site salary figures in this article. No people, quotes or statistics were invented. The "working day" is an illustrative composite.