University linear algebra, online
Linear algebra tutor online: Ajay Vatsyayan
Ajay Vatsyayan teaches university linear algebra one to one online, from row reduction and determinants to vector spaces, linear transformations, eigenvalues and the short proofs that many courses expect. Sessions follow your own course, whether it is computational, applied or proof-based, and link the ideas to where you will use them next.

1:1
Online sessions
Matrices to proofs
Computational or proof-based courses
Your syllabus
Planned from your course outline
I can row-reduce any matrix, but the moment a question says 'prove that this set is a subspace' I freeze.
Eigenvectors and change of basis feel like recipes, and my exam asks why the recipe works.
I need linear algebra properly for machine learning, quantum mechanics or differential equations, not just to pass.
Which course you are in
Computational, applied or proof-based: linear algebra comes in three styles
Two students can both be 'taking linear algebra' and be doing quite different courses. A computational course is mostly row reduction, determinants, eigenvalues and applications, often with software. An applied course adds orthogonality, least squares and dynamical systems for engineers and data scientists. A proof-based course starts from the axioms of a vector space and asks you to prove statements about bases, dimension and linear maps. The first thing Ajay does is find out which one you are in.
The official catalogues show the range. The University of Illinois offers MATH 257 Linear Algebra with Computational Applications, which includes eigenvalues and the SVD, and MATH 415 Applied Linear Algebra, aimed at engineers and covering matrices, eigenvalues, orthogonality and linear dynamical systems. Georgia Tech teaches linear algebra in the first year: MATH 1553 Introduction to Linear Algebra, MATH 1554 Linear Algebra (which adds diagonalization and quadratic forms) and MATH 1564 Linear Algebra with Abstract Vector Spaces. Purdue lists MA 26500 Linear Algebra, and UC Berkeley lists Math 56 Linear Algebra; both universities also run a combined course with differential equations (Purdue MA 26200, Berkeley Math 54).
- UK maths degrees usually meet linear algebra in Year 1 alongside analysis: Imperial's first year includes Linear Algebra and Group Theory, and Manchester lists Linear Algebra (MATH11022) as a Year 1 unit.
- Indian B.Tech students usually meet matrices, rank, eigenvalues and the Cayley–Hamilton theorem inside first-year engineering mathematics; check your university's own syllabus.
- Master's students who need the linear algebra behind a course or project can book sessions on just those ideas.
Topic map
Linear algebra topics and the exam traps in each
Most courses cover these topics in roughly this order. The right-hand column comes from the scripts we review: it is where students lose marks even when they know the method.
| Topic | The key idea | What exams ask | Trap |
|---|---|---|---|
| Systems and row reduction | Elimination to echelon and reduced echelon form | Solve a system; describe the solution set in parametric form | Arithmetic slips; calling a free variable a pivot |
| Matrix algebra and inverses | Matrices as operations that can be combined and undone | Find A⁻¹ or show A is not invertible | Assuming AB = BA |
| Determinants | Volume scaling and the test for invertibility | Compute by cofactors or row operations | Forgetting how a row swap changes the sign |
| Span and linear independence | Which vectors are redundant | Decide independence; find a vector outside a span | Checking only pairs of vectors |
| Subspaces, basis and dimension | The smallest set that builds the whole space | Prove a set is a subspace; find a basis | Not checking the zero vector or closure |
| Null space, column space, rank–nullity | What a matrix sends to zero and what it can reach | Bases for Nul A and Col A; use rank + nullity = n | Taking the column-space basis from the reduced matrix instead of A |
| Linear transformations | Maps that respect addition and scaling | Matrix of a map in given bases; kernel and image | Mixing up which basis goes on which side |
| Change of basis | The same vector described in different coordinates | Find the change-of-basis matrix and use it | Using P where P⁻¹ is needed |
| Eigenvalues and eigenvectors | Directions a matrix only stretches | Characteristic polynomial; eigenspaces | Sign errors in det(A − λI) |
| Diagonalisation | Writing A = PDP⁻¹ to compute powers | Diagonalise or explain why A cannot be | Ignoring repeated eigenvalues with too few eigenvectors |
| Orthogonality and Gram–Schmidt | Perpendicular bases make projections easy | Orthonormal basis; projection onto a subspace | Normalising too early and carrying roots |
| Least squares and SVD | Closest solution when there is no exact one | Normal equations; fit a line; singular values | Solving Ax = b directly when it has no solution |
Where students get stuck
The week linear algebra stops being about numbers
The first few weeks feel like school algebra with bigger arrays, and most students do well. Then the course introduces span, independence and subspaces, and the questions change from 'compute this' to 'show that'. Students who keep trying to compute their way through are the ones who come to us around the first midterm.
What works is learning each definition as a test you can run. 'Independent' means the only combination giving zero is the trivial one: set up that equation and solve it. 'Subspace' means three checks: zero, closure under addition, closure under scaling. Once definitions become procedures you can carry out, short proofs stop being frightening, and the computations start to mean something.
Eigenvalues are the second wall. Students can find them but cannot say what they mean, so questions on diagonalisability, repeated eigenvalues or the long-run behaviour of xₖ₊₁ = Axₖ catch them out. Ajay draws what the matrix does to a grid of points, and the algebra follows from the picture.
Worked examples
Three linear algebra problems, worked in full
Original problems at the level of a first-course midterm or final, written out the way a grader wants to see them.
Diagonalise a 2 × 2 matrix
Let A = [[4, 1], [2, 3]] (rows). Find invertible P and diagonal D with A = PDP⁻¹.
- Characteristic polynomial: det(A − λI) = (4 − λ)(3 − λ) − (1)(2) = λ² − 7λ + 10 = (λ − 5)(λ − 2). Eigenvalues 5 and 2.
- λ = 5: A − 5I = [[−1, 1], [2, −2]]. The first row gives −x + y = 0, so an eigenvector is (1, 1).
- λ = 2: A − 2I = [[2, 1], [2, 1]]. The row gives 2x + y = 0, so an eigenvector is (1, −2).
- Two distinct eigenvalues give two independent eigenvectors, so A is diagonalisable. Put the eigenvectors in the columns of P in the same order as D: P = [[1, 1], [1, −2]], D = [[5, 0], [0, 2]].
- Check: P⁻¹ = (1/3)[[2, 1], [1, −1]], and multiplying out PDP⁻¹ returns [[4, 1], [2, 3]].
Answer: P = [[1, 1], [1, −2]], D = diag(5, 2)
The order of columns in P must match the order of eigenvalues in D; swapping one without the other is a common lost mark.
Null space, rank and the rank–nullity theorem
Let B = [[1, 2, 0, 1], [2, 4, 1, 4], [3, 6, 1, 5]]. Find the rank of B, a basis for its null space, and confirm rank–nullity.
- Row reduce: R2 − 2R1 gives (0, 0, 1, 2); R3 − 3R1 gives (0, 0, 1, 2); then R3 − R2 gives a zero row.
- Reduced form: (1, 2, 0, 1) and (0, 0, 1, 2). Pivots are in columns 1 and 3, so rank B = 2.
- Free variables are x₂ = s and x₄ = t. The rows give x₁ = −2s − t and x₃ = −2t.
- So x = s(−2, 1, 0, 0) + t(−1, 0, −2, 1). The vectors (−2, 1, 0, 0) and (−1, 0, −2, 1) form a basis of Nul B; multiplying each by B gives the zero vector.
- Nullity 2. Rank + nullity = 2 + 2 = 4, the number of columns of B, as the theorem requires.
Answer: Rank 2; Nul B has basis (−2, 1, 0, 0), (−1, 0, −2, 1); 2 + 2 = 4.
A basis for the column space uses columns 1 and 3 of the original B, not of the reduced matrix.
Least squares line of best fit
Find the least squares line y = c + mt for the data points (0, 1), (1, 2), (2, 2), (3, 4).
- The system Ax = b with x = (c, m) has rows (1, tᵢ): A = [[1, 0], [1, 1], [1, 2], [1, 3]] and b = (1, 2, 2, 4). It has no exact solution.
- Normal equations AᵀAx = Aᵀb. AᵀA = [[4, 6], [6, 14]] and Aᵀb = (1 + 2 + 2 + 4, 0 + 2 + 4 + 12) = (9, 18).
- Solve 4c + 6m = 9 and 6c + 14m = 18. The determinant is 56 − 36 = 20, so c = (14·9 − 6·18)/20 = 18/20 and m = (4·18 − 6·9)/20 = 18/20.
- Check: the residual b − Ax = (0.1, 0.2, −0.7, 0.4) is orthogonal to both columns of A (its entries sum to 0, and 0·0.1 + 1·0.2 + 2·(−0.7) + 3·0.4 = 0).
Answer: y = 0.9 + 0.9t
The orthogonality check takes thirty seconds and catches almost every arithmetic slip.
Where it leads
Why linear algebra matters after the exam
Linear algebra is the course other courses quietly assume. We point out these links as we go, because seeing why a topic matters makes it easier to learn.
Machine learning and data science
Data sets are matrices; least squares is linear regression; principal component analysis and the SVD reduce dimension; gradient methods live in vector spaces.
Differential equations
A system x′ = Ax is solved with the eigenvalues and eigenvectors of A, and the solution space of a linear ODE is a vector space.
Differential equations with AjayPhysics and engineering
Quantum states are vectors and observables are operators with eigenvalues; structural, circuit and vibration problems become matrix equations.
Proof-based maths
For maths majors, linear algebra is often the first course with real proofs, and it prepares you for analysis and abstract algebra.
Real analysis with AjayHow sessions work
How we run online linear algebra sessions
Sessions take place online with handwritten working on a shared screen, at a time agreed by your time zone.
1. Send your materials
Course outline, the textbook (for example Lay, Strang or Axler, or your department's notes), current problem sheets, any returned work and your exam dates.
2. The diagnostic
Ten minutes of row reduction, then one 'show that' question. Watching both tells Ajay whether to work on fluency, on definitions, or on writing proofs.
3. Each week
The topic from your lectures: the geometric picture, the definition as a test you can run, a worked problem, then your own attempts with feedback.
4. Before midterms and finals
Timed past papers, a one-page summary of definitions and theorems you must be able to state exactly, and practice proofs in your course's style.
5. After results
We review the marked paper and agree whether to stop, continue into the next course, or work on the applications you need.
First month
A usual first month for a student who is lost at vector spaces
This is our usual plan when a student arrives around the first midterm, confident with matrices but stuck on abstraction. It is a typical approach, not a promise, and it shifts to fit your syllabus.
| Week | Work on | Goal |
|---|---|---|
| 1 | Span and independence as equations to solve | Decide independence of any set and explain the answer in one sentence |
| 2 | Subspaces, basis and dimension | Write a three-line subspace proof without prompting |
| 3 | Null space, column space, rank–nullity | Find bases for both and use the theorem to check your work |
| 4 | Linear transformations and their matrices | Build the matrix of a map in a given basis and find its kernel |
Fit
Who this is for, and who it is not for
This suits students at any university who are taking a first or second course in linear algebra, students in proof-based courses who need help writing arguments, and students in machine learning, physics or engineering who need the subject properly. Ajay holds an M.Tech and has 14+ years teaching maths.
It is not for anyone who wants graded work done for them. GATE linear algebra is taught by another Maths Bodhi tutor, not Ajay, so GATE aspirants should ask us for that tutor instead. Home tuition is available in Gurgaon (Gurugram). Students elsewhere in India and abroad learn with us online.
Further reading
Our guide on why vector spaces feel hard and how to study them.
Why linear algebra feels hardAcademic integrity
Learning support, not ghost-writing
We teach, set practice and give feedback on attempts the student has made. We do not complete graded assignments or problem sets, take online tests or exams in anyone's place, or write proofs, projects or theses that will be assessed. If you bring a graded question, Ajay teaches the method on a similar problem and you write your own answer.
Study guides
Free guides for this course
Worked examples and exam advice from the Maths Bodhi team, topic by topic.
Related Ajay routes
Continue with the most relevant next page
University maths tutor online
All the university courses Ajay teaches.
Calculus 3 tutor online
Multivariable and vector calculus.
US college calculus tutor
Calculus I to III and the US course sequence.
IIT first-year maths
Calculus and linear algebra in the first year at IIT.
Engineering mathematics tutor
B.Tech M1, M2 and M3, including matrices.
Differential equations tutor online
Where eigenvalues meet ODE systems.
Book a free demo class
Bring your hardest linear algebra question.
Ajay Vatsyayan
Profile and the full list of courses.
FAQ
Questions students ask about linear algebra tutoring
Can you help with linear algebra proofs?
Yes. We work on the definitions and standard arguments, such as subspace tests, independence and dimension proofs, until you can write them yourself.
My course is computational with MATLAB or Python. Is that covered?
Yes for the mathematics behind the code: row reduction, eigenvalues, orthogonality, least squares and SVD. Sessions focus on understanding, not on writing your graded code.
Which textbook do you follow?
The one your course uses. Send the chapter list and problem sets and we keep your notation.
Is linear algebra harder than calculus?
It is different. The calculations are easier, but abstraction arrives sooner, and that is where most students need help.
I need linear algebra for machine learning. Can we focus on that?
Yes. We can concentrate on vectors, matrix products, projections, eigenvalues and the SVD, with the reasons each matters for models.
Do you teach students outside India?
Yes. Sessions are online and times are agreed by your time zone when you enquire.
How many sessions does a student usually need?
Many book one or two a week through a semester. A student close to an exam may book more for a few weeks.
Is matrices in B.Tech M1 the same as university linear algebra?
It overlaps. Most B.Tech first-year courses cover rank, systems, eigenvalues and Cayley–Hamilton, while vector spaces and proofs are lighter. Send your syllabus and we match it.
Will you do my problem set?
No. We teach the method and review your own attempts; we never complete graded work.
Next step
Share your linear algebra course
Send your university, course name, textbook, a recent problem sheet and your time zone. Maths Bodhi will check whether Ajay is the right fit and suggest session times.
