Walk into any engineering college placement cell in India right now and you will hear some version of the same question: "I want to get into data science — what do I do?" The trouble is that "data science" in 2026 covers four genuinely different career tracks, each with different skill requirements, different salary bands, and different hiring pipelines. Conflating them is how freshers end up spending a year learning the wrong things and then wondering why the job market feels impenetrable.
This roadmap is built around the Indian job market as it actually exists — not as it was three years ago during the peak hiring boom. The demand is real, but it is more specific than the generic "become a data scientist" content online would suggest. Knowing which track you are targeting makes the difference between a focused 12-month plan and an endless loop of tutorials.
The Four Data Roles in the Indian Market — And How They Differ
The Indian data job market has split into four distinct roles with meaningfully different day-to-day work, tool stacks, and hiring criteria. Choosing the right one early saves months of misdirected effort.
Data Analyst
Data analysts answer business questions using existing data. The core tools are SQL, Excel, and a BI tool (Power BI or Looker Studio in most Indian companies, Tableau in larger firms and GCCs). The role requires business context — understanding why a metric matters, not just how to calculate it. This track has the most entry-level jobs in India, particularly at e-commerce companies, startups, and traditional enterprises building out their analytics function. Salaries for freshers in Bangalore start at ₹4–7 LPA. Less glamorous than ML, but far easier to get that first job.
Data Engineer
Data engineers build and maintain the pipelines that move, transform, and store data. The tools are SQL, Python, dbt, Airflow, and cloud platforms (BigQuery, AWS Redshift, or Azure Synapse depending on the company's stack). This is currently the highest-growth track in India — GCCs in Hyderabad and Bangalore are hiring aggressively, and the supply of qualified data engineers is still short of demand. Entry-level salaries at product companies run ₹7–12 LPA. The work is less about statistics and more about engineering reliability and scale.
ML Engineer
ML engineers take models from notebooks into production — they handle deployment, monitoring, retraining pipelines, and inference infrastructure. This role requires strong Python coding skills, familiarity with MLOps tools (MLflow, Weights & Biases, Kubeflow), and comfort with Docker and cloud deployment. It is the most technically demanding entry-level track and the most selective. Companies like Flipkart, Swiggy, and the Indian arms of Google and Microsoft hire ML engineers, but the interview process is closer to software engineering than data analysis.
Research Scientist
Research scientists work on novel model development — mostly in academia, large AI labs, and a handful of well-funded Indian AI startups. This track is the most selective and almost always requires a Masters or PhD. For a fresher without postgraduate credentials, this path is a multi-year journey, not a 12-month sprint. It is the smallest of the four segments in terms of job volume in India.
Months 1–3: Building the Foundation Everyone Actually Needs
Regardless of which track you eventually target, three skills form the foundation: Python basics, data manipulation with pandas and NumPy, and data visualisation with matplotlib or seaborn. These are non-negotiable entry requirements across all four tracks.
Python basics means being comfortable with data types, loops, functions, list comprehensions, and file I/O — not object-oriented programming or decorators. Start with the official Python tutorial, supplement with Automate the Boring Stuff with Python (free online), and then move directly into pandas. Spending three months on pure Python theory before touching data is a common mistake — get to pandas within four weeks.
The pandas library is where most freshers underinvest. Go beyond reading and filtering DataFrames. Learn groupby aggregations, merge and join operations, pivot tables, and how to handle missing data. The Analytics Vidhya learning path for pandas is one of the better structured free resources in the Indian ecosystem. Supplement with Kaggle's free micro-courses, which are practical and well-paced.
By the end of month 3, you should be able to load a CSV file, clean the data, answer five business questions from it, and produce a clear chart for each answer. That is the concrete milestone. If you can do that fluently, the foundation is solid.
Months 4–6: SQL and Business Context — The Underrated Middle Phase
SQL is the single most underrated skill in the Indian fresher data ecosystem. It is tested in nearly every data interview across all four tracks. Analysts use it daily. Engineers build on top of it. ML engineers query feature stores with it. And yet most online courses treat it as an afterthought after machine learning.
Spend two months becoming genuinely proficient in SQL — not just SELECT and WHERE, but window functions, CTEs, subqueries, and date manipulations. The PostgreSQL dialect is the most transferable; learn it first and BigQuery SQL second if you are targeting cloud data roles. SQLZoo, Mode Analytics' free SQL tutorial, and LeetCode's database problems (easy and medium) are the right practice grounds.
The business context layer matters as much as the syntax. Learn what DAU, MAU, retention rate, cohort analysis, and funnel conversion mean in product analytics. Understand how to read a basic P&L. Read case studies from Indian companies — Flipkart's data blog, PhonePe's engineering blog — to see how real analysts frame business problems. This context is what separates candidates who can query data from candidates who can answer the question the business is actually asking.
In month 6, add GA4 fundamentals and basic statistics: distributions, hypothesis testing, A/B test interpretation. These appear in analyst interviews at every Indian product company.
Months 7–9: Specialise by Your Chosen Track
By month 7, you should have a clear sense of which track fits your strengths and goals. Here is what to prioritise in each:
Analyst Track
Build fluency in one BI tool — Power BI is the safest choice for the Indian market because Microsoft's enterprise dominance means most mid-size companies use it. Learn DAX basics, data modelling in Power BI Desktop, and how to publish reports to Power BI Service. Supplement with Looker Studio for startups that prefer Google's ecosystem. Tableau is worth learning only if you are specifically targeting large enterprises or MNC GCCs where Tableau licenses are already in place.
Data Engineer Track
Start with dbt (data build tool) — it has become the de facto standard for SQL transformations in modern data stacks, and Indian companies adopting the modern data stack are specifically looking for dbt skills. Then add Apache Airflow for orchestration. For cloud, BigQuery is the most practical starting point because it has a generous free tier for learning and is widely used in Indian product companies. The dbt Learn platform offers free courses that are genuinely good.
ML Engineer Track
Focus on scikit-learn for model building, but spend equal time on feature engineering and model evaluation — these are what interviews actually test. Learn how to use MLflow for experiment tracking (it is open-source and free). Understand how to package a model as a REST API using FastAPI, and how to containerise it with Docker. The Practical Deep Learning for Coders course by fast.ai is an excellent free resource for going deeper into neural networks if that is your interest.
Building a Portfolio That Gets Responses in India
The portfolio problem for Indian freshers is that most online project tutorials use the same three datasets — Titanic, Iris, Boston Housing — and hiring managers in India see hundreds of these. A portfolio that stands out uses Indian data and answers questions that are meaningful to Indian businesses.
Three sources of Indian datasets worth knowing:
- data.gov.in — India's open government data portal has datasets on agriculture, health, transport, and demographics across all states. Building an analysis on crop yield data by district or hospital bed availability by state shows genuine initiative.
- CMIE (Centre for Monitoring Indian Economy) — their Prowess database tracks listed Indian companies. Free access is limited, but the publicly released unemployment and GDP data are useful for macro-level analysis projects.
- NSE and BSE — stock price data for Indian companies is freely available via the NSEpy Python library and the official NSE CSV downloads. A portfolio project that analyses sector rotation or earnings surprise effects has immediate relevance to Indian fintech and wealth management recruiters.
Two to three well-documented projects beat ten shallow ones. Each project should have a GitHub repository with a clear README explaining the business question, the data source, the approach, and the finding — not just the code. Recruiters at Indian companies who look at GitHub want to see that you can communicate results, not just run notebooks.
Internships via Internshala are worth pursuing seriously during months 7–9. Even a two-month paid data internship at a small Indian startup adds more credibility to a resume than two additional certifications. Companies like GreytHR, Zoho, and the analytics teams at various edtech firms regularly hire interns through Internshala with real project work.
Months 10–12: Applying Strategically to Indian Companies
The Indian data job market has a clear hiring hierarchy for freshers. Applying to Google India or Walmart Global Tech as your first application is a calibration error — use those applications as benchmarks after you have practiced on reachable targets.
The realistic hiring pipeline for a fresher with a solid portfolio looks like this:
- Analytics service companies (Mu Sigma, Fractal Analytics, Tiger Analytics, LatentView) — these companies hire freshers in volume, the work is client-facing analytics, and they provide training. Salaries are lower (₹4–6 LPA) but the learning is fast and resume value is real.
- Mid-tier product startups (Series B and C companies in edtech, fintech, D2C e-commerce) — they often lack a structured data team, which means more ownership and faster growth. Look for companies in the ₹50–500 crore revenue range with a data or analytics job posting.
- GCCs (Global Capability Centres) in Bangalore and Hyderabad — companies like Walmart, Target, Lowe's, and JPMorgan Chase run large data teams out of India. They hire freshers, the work is close to what the parent company does, and salaries are competitive at ₹7–12 LPA for data engineering roles.
- Large Indian product companies (Flipkart, Swiggy, Zomato, PhonePe, Razorpay) — hardest to get into without prior experience, but not impossible with a strong portfolio and referral. Referrals matter enormously in Indian tech hiring.
For the interview process: expect a take-home SQL assignment, a Python data manipulation task, and a business case discussion in analyst interviews. ML engineer interviews add a coding round closer to software engineering interviews (LeetCode medium difficulty) plus an ML system design round. Prepare for at least three to four rounds before an offer.
Frequently Asked Questions
Which pays better in India in 2026 — data engineering or data science?
Data engineering consistently pays 15–25% more than data science at entry level in India right now. The gap exists because demand for data engineers is outpacing supply, while the data science job market is more competitive due to the surge in bootcamp graduates. A fresher data engineer in Bangalore can expect ₹7–10 LPA at a mid-tier product company, while a data science fresher at a comparable company is more likely to start at ₹5–8 LPA. The salary trajectory also diverges — senior data engineers at large tech companies command ₹30–50 LPA, while senior data scientists without strong ML deployment skills often plateau earlier.
Do I need a Masters degree to get a data science job in India?
For data analyst and data engineer roles, no — a strong portfolio and demonstrable SQL and Python skills matter far more than a postgraduate degree. Most hiring managers at Indian product companies (Flipkart, Swiggy, PhonePe, CRED) evaluate candidates on take-home assignments and coding rounds, not degrees. For ML engineer roles at companies like Google India, Microsoft India, or research-focused firms, a Masters or equivalent rigour is helpful but not an absolute gate. Where a Masters genuinely helps is in differentiating you when you have no prior work experience and your portfolio is thin.
Which Indian cities have the most data science job opportunities in 2026?
Bangalore accounts for roughly 55–60% of all data roles in India — it is where the product companies, funded startups, and Indian R&D centres of global tech firms concentrate their data teams. Hyderabad is a solid second, particularly for roles at Microsoft, Amazon, and the large GCC operations of US companies. Mumbai is the right city for fintech data roles — companies like CRED, Zepto, and traditional financial institutions are building serious data teams there. Pune is growing fast, especially for data engineering roles at product companies. For someone in Kerala specifically, most serious data roles require relocation or remote arrangements, though Kochi is seeing gradual growth in GCC data hiring.