Interview Question Bank (Tech)
Interview Question Bank (Tech)
Linear Algebra
1. Eigenvalue definition
Eigenvalue is the scale factor by which the eigenvector is stretched
2. Eigenvector definition
A special vector that remains at its own span when multiplied by a matrix system. Basically,
if you multiply it with the matrix, you would get a scaled version of that original vector.
3. Use of Eigenvectors
Is used in some advanced techniques to reduce noise in systems. Also, I remember a specific
example of using the eigenvector to find the axis of rotation of a 3D Shape.
4. Diagonal Matrix and Eigenvalues
If your matrix is a diagonal matrix, all the basis vectors are eigenvectors & the diagonal
entries are the eigenvalues.
5. Eigen-functions
Eigenfunction is a function that returns the original input times a constant factor when
applied. Example: The derivative of Exponential.
6. The equation for finding eigenvalues det. ( A – lambda I) = 0
To find the Eigenvalues, we use the characteristic equation
7. Basis and Subspaces, vector space, Eigensystem, eigenspaces
Basis are a set of linearly independent vectors that can span all other vectors in a Vector
space.
Dimension is the number of vectors in the basis of a vector space.
The basis can vary, but the dimension size would always be the same.
Vector space is a set of vectors that abide by all properties of linear vectors.
Eigensystem is simply a pairing of Eigenvalues & their corresponding eigenvectors
Eigenspace is a set of all eigenvectors of a system
8. How to solve system of linear equations.
Even though we can solve a system of equations graphically, but on the linear algebra side,
we can use the matrix approach and use Gauss Jordan or Gauss Elimination to bring our
system into a row echelon form to solve it.
In row echelon form, there is a leading entry and all entries before that one is zero.
9. What are under-determined, over-determined systems
Underdetermined systems are those which have fewer linearly independent equations than
the unknowns.
Overdetermined systems are those which have more linearly independent equations than
the unknowns.
10. Rank of a matrix
Rank is the number of linearly independent rows / columns of a matrix. Basically it gives an
idea if a matrix system is solveable.
We can find the rank by bringing the matrix into a row echelon form and then checking he
non-zero rows.
11. Determinants
Determinant is a special scalar number that can be calculated from a matrix. If the
determinant is non-zero, the matrix is “full rank” which means the number of rows are all
linearly independent.
12. How will you solve a 100 x 100 Matrix?
I believe if I can use an online calculator to evaluate the determinants, the Cramers rule
could be the best option!
13. Orthogonality
Two vectors are said to be orthogonal if they have none of their components over each
other. Basically, their dot product is zero. All orthogonal vectors are by default linearly
independent. However, it is not true in the other way.
A matrix is orthogonal if its transpose is equal to its inverse.
Probability
1. Gaussian, Normal Curve and its controlling parameters
Gaussian Distribution is also known as the Normal distribution curve. Basically, the feature
of curve is that it is symmetric about the center. Which means this distribution is showing
the fact that the data near the mean occurs more frequently.
The two controlling parameters of a gaussian distribution are the mean & the standard
deviation.
2. A random variable, various pdfs and marginal pdfs etc
Random variable is a measurable quantity in any chance process. For instance, I toss a coin
10 times. The number of heads I get could be the random variable.
Probability density function is specifically for a continuous random variable and it shows
the likelihood of an outcome being close to a specific value. Area under the graph gives the
probability.
Types include Chi-Square, Normal, Poisson & Binomial etc. Each distribution represents a
different scenario or case.
Marginal Distribution is a subset of the probability function
3. Probability: what is conditional probability, bays theorem?
Conditional Probability basically tells you the chance of A happening given that B has
already happened.
Bayes theorem is a statistical relation that can link one conditional probability with its
inversion. For instance, if I know the probability of A happening given B has happened,
Bayes theorem would allow me to find the probability of B happening given A has
happened.
4. What is the formula for a Gaussian curve assuming variance =1?
It would be 1 by 2pi into exp of (-1/2) times (x-mean)squared
5. pdf, cdf, pmf
Probability Distribution function gives the probability distribution of a continuous random
variable. Gives probability of a value being in a given range.
Probability Mass Function gives the probability distribution of a discrete random variable.
Cumulative Distribution Function is applicable to both continuous and discrete function.
Basically, CDF simply tells the probability of a random variable being less than or equal to
a specific value.
6. Name some discrete or continuous distributions.
Examples of Discrete distributions include Poisson & Binomial.
Examples of Continuous distributions include Gaussian, Chi-Square and Uniform.
7. Gaussian and binomial distribution.
Gaussian distribution is a type of continuous probability distribution. It basically assumed
that the data centered around the mean is the most likely to occur. This produces a bell-
shaped curve.
Binomial distribution is a type of discrete probability distribution. It is basically used to
model a specific number of “successes” in an experiment that is repeated a fixed number of
times.
8. What is law of large numbers?
The law of large numbers states that if an experiment is repeated a large number of times,
the experimental probability would start to approach the expected i.e theoretical
probability.
Communication Systems
1. What is Source Coding, Channel Coding, and Line Coding?
Source coding is basically a form of encryption. We remove the redundancy from our
databits in order to reduce the bandwidth requirement of the signal. This compresses our
digital signal.
Channel Coding comes after Source coding. We add the error checking into the data, for
example parity checks etc.
Line coding is the last stage where the data is finally converted into voltage signals/pulses.
2. What is CDMA?
Channel Division Multiple Access is a spectrum-based technology where each unit is
assigned a code and then they are multiplexed over the same channel for multiple access.
3. What is the Cut-Off Frequency?
It is the frequency at which the response is -3Db with respect to maximum response.
4. What is Antenna?
Antenna is basically a transition device between a guided wave & free space. It is used to
transmit or receive wireless signals.
5. Baud Rate
The rate at which bits are transferred in a digital channel.
6. BPSK definition
BPSK is a type of digital modulation where the phase of the carrier is switched in two
variants based on the bit of the message signal.
7. QPSK definition
Quadrature Phase Shift Keying allows for the phase of the carrier being keyed in 4 different
values. It has an advantage over the BPSK as it allows for double data rate while
maintaining the same bandwidth.
8. Difference b/w FM and AM.
In Frequency Modulation, the carrier’s frequency is altered/modulated with the message
signal.
In Amplitude Modulation, the carrier’s amplitude is altered/modulated with the message
signal.
AM Is more susceptible to noise while FM isn’t. However, AM is easy to implement while
FM is a little complex.
9. Different digital shift keying or modulation schemes
10. Difference b/w TDM and FDM
TDM stands for Time Division Multiplexing. In this type of Multiplexing, the time sharing
takes place. One information is sent at one time, the next at another time etc. TDM works
for both analog and digital signals.
FDM stands for Frequency Division Multiplexing. In this type of multiplexing, the
frequency band is divided and data is transmitted simultaneously at different bands. This
works only for analog signals.
11. Advantage/Disadvantage of TDM over FD
TDM is easier to implement than FDM. Also, there is less conflict than FDM as there is no
interference of bands.
In TDM we do require a synchronization pulse.
Latency is also low with TDM as each signal is sent after the previous one.
12. What is the function of equalizer?
An equalizer is essentially used to alter the frequency response of a signal. There are
multiple cases of an equalizer being used, for instance, to simply adjust the parameters or
the tune of music. However, in wireless communications, it is also used to reduce distortion
or get a correct signal back.
13. Name an equalizer used for TDMA?
Adaptive Equalizer
14. Can we convert a bandlimited signal to digital? How?
Sample the signal at a frequency which is atleast twice than the highest frequency
component. Sampling basically means you multiple it with an impulse function.
15. What are Error Detection/Correction codes?
Error Detection codes are appended to the original bitstream and they simply tell the
receiver if they have gotten the correct data from the transmitter. Parity check is an
example of such a code.
Error Correction codes are appended to the original bitstream and they are actually a step
ahead of detection codes. They can also allow the receiver to fix any errors in the bitstream.
16. Main classes of Error Detection/Correction codes?
Error Detection: Parity Check, Cyclic Redundancy Check (CRC), Checksum
Error Correction: Hamming Code, Automatic Repeat Request (ARQ)
17. Give some examples of Error Detection/Correction codes
18. Error detection or Correction codes. Difference b/w them.
19. LAN/WAN
LAN is a Local Area Network. It is basically in a confined space where different computers
or devices are connected together in a closed network.
WAN is Wide Area Network. It is a bigger scale of a Local Area Network. Multiple LANs
connected together can form a WAN. Internet is an example of a massive WAN.
Electronics
1. Define CMOS
CMOS stands for Complimentary Symmetry Metal Oxide Semiconductor. Basically, it is a
symmetric mix of the NMOS & PMOS. The main advantage of these are their high noise
immunity and very low power dissipation.
The CMOS basically only consume external power in the switching states. This allows for
less consumption of heat than other logic circuits.
Use cases include microcontroller integrated chips.
2. Advantage of NAND over NOR
NAND gate is generally more robust than NOR gates.
3. What is Fan-Out
Number of Digital inputs that a single output can feed. TTL has a lower fanout than
CMOS.
4. What are universal gates? Advantage of them?
The NAND and NOR gates are known as universal gates as they can implement all
combinational and sequential circuits. For instance, a NAND gate can also be used to
construct a flip flop.
5. What is the Wheatstone bridge?
The wheatstone bridge allows for a full bridge measurement with 3 resistors and one sensor.
It allows for a high Common Mode Rejection ratio due to it being a null measurement.
6. What is an Op-Amp?
Op Amp stands for Operational Amplifier. Basically, it is an electronic device that is used to
scale an input voltage and essentially amplify it.
Ideal Op-Amp properties: Infinite Input resistance, Zero output resistance, Infinite CMRR,
Infinite Bandwidth
7. What is Common Mode Rejection Ratio
This is the ability of a device to reject common mode signals. Signal that are common on
both terminals of inputs such as noise.
8. Diodes & Types of diodes
Diode is a two terminal device where current can only be conducted in one direction.
Types include: P-N junction diode, Schottky Diode, Zener Diode
In Schottky diode, the forward voltage is very low so that allows for high speed switching.
Zener diode also allows for current to be conducted in reverse direction when a certain
voltage threshold is reached.
9. Difference in working of BJT & CMOS
BJT stands for Bipolar Junction Transistors. They basically contain both electrons & holes
as charge carriers.
MOSFET instead has only one charge carrier.
MOSFET only consumes external power in switching mode so hence, power dissipation is
much lower. Also, FETs are more immune to noise.
Intermediate resistance between on and off state in MOSFET is significant, hence they can
sometimes dissipate large energy in switching state.
10. Regions of CMOS
There are three regions if I remember currently. Cut-Off region, Triode Reigion &
Saturation Region. The cut-off and triode regions are used to operate as a switch. The
saturation region is used to operate as an amplifier.
11. Difference between PMOS & NMOS
PMOS has holes as charge carriers, NMOS has electrons as charge carriers.
In PMOS, we need active low on base to activate, in NMOS we need active high.
NMOS is generally more quicker in switching as electrons travel more quickly than holes.
However, PMOS is less susceptible to noise.
12. Define 555 Timer IC and its usecases
The 555 timer IC is used for timing applications such as pulse generation or oscillators.
They can basically be used to provide time delays and even as flip flop elements.
Three Modes: Monostable, Astable, Bistable.
13. No. of Transistor used in constructing AND, OR gates using CMOS
2 transistors are used to make AND, OR / NAND gates. For the NOR gate, we can use 1 or
2.
14. Power dissipation in BJT and CMOS. Which one in better? Difference b/w CMOS and Bipolar
technologies.
CMOS dissipates much less power than the BJT. The reason is, for a BJT to switch on, we
need current flowing into the base. However, in the MOSFET technology, there must only
be a voltage at the base. This means that the CMOS does not dissipate as much power in a
static state when it is not switching.
The key difference is BJT modulates a current by varying the input (base) current while the
MOSFET modulates a current by varying the input (base) voltage.
15. If you have a circuit with resistors, inductors and capacitors, how would you find the
voltage/current at different points (nodal/mesh analysis)
The easiest way to process such a circuit would be to use laplace transform to convert the
components and then use a simple linear analysis.
16. How do you decide whether or not to perform nodal or mesh analysis?
Depends on the number of nodes, number of meshes also the number and type of sources.
17. How will you obtain the characteristic graph of a RC circuit?
Connect a resistor in series with a capacitor. Turn on the battery and monitor voltage
across the capacitor with a voltmeter. This will give the charging curve.
For the discharge curve, disconnect battery and replace with short circuit. The capacitor
will discharge through resistor.
Programming
1. Do you know C programming?
2. Why is java so easily compatible with any system?
3. Different layers in an OSI networking model.
Application layer at the top.
Presentation Layer.
Session Layer.
Transport Layer.
Network Layer.
Datalink Layer.
Physical Layer.
4. Switches
Switch is a device that forwards data packets to a specific device based on a MAC address.
5. What is size of integer
It actually depends on the language. In C, an integer is worth 2 bytes. In java, I believe its 4
bytes. In LabVIEW, they combine int with double int and other forms and give various
alternatives together.
6. What is MATLAB used for
MATLAB has a variety of uses. Some include modelling systems, performing advanced
mathematical operations and even developing advanced algorithms.
7. Different plot features in MATLAB
Box plot, Hold feature, Plot, Plot3 (3D), Xlim etc
8. What are XMATH features
Debugging tool with GUI Features
9. How to solve ODEs in MATLAB
There are functions such as Quad, Quad8, ODE23 & ODE45.
10. How to add noise in MATLAB
There is an inbuilt Noise mscript file that is used to add various types of noises like gaussian
and normal etc.
11. Why OOP
12. Datastructures
https://emerginged.com/blogs/emerginged-blog/eight-data-structures-every-programmer-
has-to-know?
ga_client_id=GA1.2.1255236468.1647172658&viewId=emergingedl0p82m5acfq3t5