I have a dataset where lot of names are written like man1sh instead of manish, vikas as v1kas.
How can one correct these names in NLP?
The errors you describe are typically for scanned text and are introduced by the imperfections of Optical Character Recognition (OCR).
Search for "OCR correction" to get an overview of techniques available to correct those errors.
For highest quality, manual correction (with at least two independent proofreaders) is still the best we have.