At night, I went outside. I had a box with two slits in it. I directed torch light towards it, but I saw only two bands of light on the wall and shadow from the rest of the box. Why did it not produce interference like a double slit experiment should?
In order to see the interference fringes, four conditions must be fulfilled:
Your light source either has to be point-like or very far away from the slits,
Your light source must be monochromatic* (i.e. emit only at a single wavelength),
The slits must be very close together, and
The slits must be very thin.
Failing to meet any of these will generate enough noise to completely obscure any signal you hope to measure. Probably the first one to fix in your case is to switch from a torch (which is neither point-like nor monochromatic) to a laser (which is much closer to being both of those things).
*In order to see the usual interference fringes, you must use monochromatic light. There is a similar effect you can do with a point source of white light, though, as shown in the comments.
The other answers are correct, you need a coherent, monochromatic light source to see interference fringes. This site gives detailed instructions on how to do it using a laser pointer and a nit comb:
You can buy everything you need on ebay. You will likely pay less for the laser pointer than the nit comb! (Make sure it really is a laser. When I searched I got lots of hits for LED cat toys. Not the same thing).
First of all, you need a laser, which is a reliable source of coherent light. The second thing you should know, to get the interference pattern, the dimensions of the slits (or lets say apertures, they don't need to be slits) should be comparable to the wavelength of the light you are using.
Let's say you are using a good old red laser pointer, which has a wavelength about 632 nm (visible region is ~350-700 nm). You need to have apertures with dimensions about a few micrometers (same for other visible region colours). You could observe interference pattern with two tiny circular aperture or rectangular ones. You can use whatever shape you like, although the pattern you observe will have different shapes according to the different aperture shapes. What is important is that at least one dimension of the apertures should be about a few micro meters.
Note: You could also see the interference pattern for polychromatic light sources but it would be smudged as different colours of light will interfere on different locations on the screen.