I started training from mostafa.saad.fci's junior sheet. But as I'm busy now studying for my faculty and other things I am currently participating in contests only and up solving one or more problems after these contests.

Figure out where the main issue lies. Is it the implementation you generally get stuck with, or is it challenging to get ideas about solving the problem?

If it's the implementation, you should look at others' solutions after solving the problem. This will help you know different ways in which you can implement the solution efficiently.

For the second one: 1) just upsolve 2) give virtual contests

Here is how I like to practice after a contest, based on my own experience and the advice of some ICPC Finalists.

Let's call "problem N" the last problem you solved during a contest. Then "problem N+1" is the first you couldn't solve, "problem N+2" is the one after that, etc.

After a contest, you should focus on problems N, N+1 and N+2. Problem N-1 is too easy and not worth spending more time on, and problem N+3 is too hard and outside of your league.

Reviewing a solution

After getting "Accepted" verdict, it is worth taking the time to reflect on the solution, how you could come up with it, and whether there is a simpler way.

To accomplish this you may (in this order):

1. Explicitly say all the insights and observations needed to solve the problem, out loud.
2. Think about using simpler algorithms and data structures (e.g. I used segment tree for this problem... Would partial sums work too?).
3. Try to find observations that will simplify your code.
4. Read the editorial and think about how it compares to your own solution.
5. Look at solutions by top competitors, try to figure out the idea and implement it yourself (without looking at their code again).

Problem N

You already know how to solve it, but you might be able to simplify your solution.

After the contest, jump straight to review. Focus particularly on steps 1, 2 and 3.

Problem N+1

You probably know enough techniques to solve it, but couldn't figure out the logic within the contest time.

Given enough time, you will probably be able to solve it. So try really hard for 2 days (or 3 if you feel you're close).

If you fail, read the editorial and implement the solution.

After getting "Accepted", go into review. Focus on steps 1, 3 and 5.

Problem N+2

This problem is probably too hard for you to get something out of on your own.

Read the editorial right away, and implement the solution. Then, you can go into review, focusing on steps 3 and 5.

You need to be comfortable doing B problems. So you can also starting solving div2 B of past contests. I think they are in 1000-1200 range mostly