Quick answer. DSIM is the Department of Statistics and Information Management cadre of the Reserve Bank of India. Selection is through three papers worth 300 marks plus a 75 mark interview, giving a final merit out of 375. Paper I is objective Statistics for 120 minutes, Paper II is descriptive Statistics written by hand for 180 minutes, and Paper III is English typed on a keyboard for 90 minutes. The syllabus has nine units and now formally includes machine learning, neural networks and SQL. The 2026 cycle had 10 vacancies.
At a Glance
| Particular | Details |
|---|---|
| Cadre | Department of Statistics and Information Management |
| Vacancies, 2026 cycle | 10 |
| Papers | 3 |
| Written total | 300 marks |
| Interview | 75 marks |
| Final merit out of | 375 |
| Syllabus units | 9 |
| Age limit | 21 to 30 years |
| Minimum qualification | Master’s with 55 percent, or 4 year Bachelor’s with 60 percent |
| Starting basic pay | Rs 78,450 per month |
| Calculator | On screen non programmable, Paper II only |
The Three Papers
| Paper | Type | Duration | Marks |
|---|---|---|---|
| Paper I | Objective, Statistics, online | 120 minutes | 100 |
| Paper II | Descriptive, Statistics. Questions on screen, answers written on paper | 180 minutes | 100 |
| Paper III | English, descriptive, typed on a keyboard | 90 minutes | 100 |
Paper I is held on one date. Papers II and III are held together on a second date, several weeks later.
Only candidates who clear the Paper I cutoff have Papers II and III evaluated. The cutoff is decided by the Board in relation to the number of vacancies.
But Paper I marks still count in the final merit. This is not a screening test that is then discarded. All three papers plus the interview are added together. A candidate who scrapes through Paper I carries that deficit all the way to the end.
The Nine Syllabus Units
Both Paper I and Paper II draw from the same nine units.
| Unit | Title |
|---|---|
| 1 | Theory of Probability, Probability Distributions and Sampling Theory |
| 2 | Linear Models and Economic Statistics |
| 3 | Statistical Inference: Estimation, Testing of Hypothesis and Non Parametric Tests |
| 4 | Stochastic Processes |
| 5 | Multivariate Analysis |
| 6 | Econometrics and Time Series |
| 7 | Optimization and Statistical Computing |
| 8 | Data Science, Artificial Intelligence and Machine Learning Techniques |
| 9 | Database and Data Warehouse Management |
Units 8 and 9 are the story here. Neural networks, random forests, gradient boosting, natural language processing, SQL joins, star and snowflake schemas and OLAP versus OLTP are now formally inside a central banking examination syllabus.
The Board also states plainly that the standard of the papers is that of a postgraduate degree examination.
Read the full nine unit syllabus
What the Job Actually Involves
Worth reading, because the interview tests whether you understand it.
The official job description says the role involves collecting, compiling, analysing and interpreting data on the banking, corporate and external sectors, with heavy use of statistical and econometric techniques to model and forecast inflation and growth. It names classical and Bayesian inference, high dimensional regression, time series models and simulation based methods.
It also names designing and executing complex sample surveys, understanding economic statistics, index number construction and inequality measurement, managing large structured and unstructured datasets, writing and optimising database queries, and applying data science, artificial intelligence and machine learning to extract insights.
That is a data science role inside a central bank, and the syllabus reflects it exactly.
Eligibility
Any one of three routes:
- Master’s with 55 percent in Statistics, Mathematics, Mathematical Statistics, Applied Statistics, Quantitative Economics, Econometrics or Informatics
- Master’s with 55 percent in Data Science, Artificial Intelligence, Machine Learning or Big Data Analytics
- Four year Bachelor’s with 60 percent in any of the above
SC, ST and PwBD candidates need 50 percent, where posts are reserved.
Age 21 to 30 years, with 5 years relaxation for SC and ST and 3 years for OBC.
Full eligibility, fee and vacancy details
The Previous Year Paper Problem
The most searched phrase among DSIM aspirants is “previous year papers PDF”, and the honest answer is that not enough of them exist to build a preparation around.
That feels like a disadvantage. It is not, once you see what it implies.
Across recent cycles the Paper I cutoff for the general category has moved across a wide range. The difficulty swings sharply from year to year. A candidate who prepared by memorising patterns from two memory based papers is preparing for the wrong exam.
Concepts are the new previous year questions. A question on the Cramer Rao inequality, ridge regression or a Markov chain can be framed in a dozen ways, and the underlying idea does not move. Build at that layer and the shortage stops being a weakness.
Read the full concept first roadmap
Where to Borrow Practice From
Since DSIM specific papers are limited, take rigour from examinations that test the same concepts.
| Source | What it gives you |
|---|---|
| UPSC ISS Statistics papers | The closest cousin of DSIM Paper II. Long form descriptive answers on the same topics. |
| IIT JAM Mathematical Statistics | Clean objective practice on probability, distributions and inference |
| GATE Statistics | Objective practice at a demanding level, with the same core units |
| UGC NET Statistics | Broad coverage including sampling, design of experiments and time series |
| ISI entrance and CUET PG Statistics | Additional objective and analytical practice |
UPSC ISS is the most valuable of these and it is the one most DSIM aspirants overlook. Its descriptive Statistics papers are the nearest available rehearsal for writing DSIM Paper II by hand.
Books
A short list studied deeply beats a long list studied once.
How to Approach It
| Stage | Focus |
|---|---|
| 1 | Build the classical Statistics core: Units 1, 2, 3 |
| 2 | Add the depth units: 4, 5, 6 |
| 3 | Optimization and Statistical Computing, Unit 7 |
| 4 | Units 8 and 9, Data Science and Databases. New for almost everyone. |
| 5 | Descriptive answer writing practice, by hand, using ISS papers |
| 6 | Paper III English: precis, essay, comprehension, typed |
| Throughout | Objective practice under time pressure |
Do not leave Units 8 and 9 to the end. They are the units most Statistics postgraduates have never studied, and they are the ones that decide who can answer across the whole paper rather than half of it.
Frequently Asked Questions
What is RBI Grade B DSIM?
DSIM is the Department of Statistics and Information Management cadre of the Reserve Bank of India, recruited through Grade B direct recruitment.
How many papers are there in RBI DSIM?
Three. Paper I objective Statistics for 120 minutes, Paper II descriptive Statistics for 180 minutes written by hand, and Paper III English typed for 90 minutes. Each carries 100 marks.
How many marks is the RBI DSIM interview?
75 marks, taken together with the 300 marks from the three papers for a final merit out of 375.
How many units are in the RBI DSIM syllabus?
Nine, covering everything from classical probability to machine learning and database management.
Is machine learning really in the RBI DSIM syllabus?
Yes. Unit 8 explicitly includes logistic and penalized regression, naive Bayes, nearest neighbour, decision trees, support vector machines, neural networks, random forests, boosting, recurrent and convolutional neural networks and natural language processing.
Is a calculator allowed in RBI DSIM?
Physical calculators are not permitted in any phase. A non programmable calculator is provided on screen for Paper II only, and only for DSIM candidates.
How many DSIM vacancies are there?
The 2026 cycle advertised 10 DSIM vacancies out of 60 total Grade B posts.
Are previous year papers available for RBI DSIM?
Only a limited number of memory based papers exist. This is why a concept first approach, supplemented with practice borrowed from UPSC ISS, IIT JAM, GATE and UGC NET Statistics, works better than pattern hunting.