Reducing granularity(from 8 to 4) along with lowering the false positive rate(0.01 to 0.001) was effective in skipping unnecessary granule reads. Setting false positive rate to 0.0001 was giving more accuracy but with bigger bloom filter size and memory usage.
ALTER TABLE testtable ADD INDEX testindex (testColumn) TYPE bloom_filter(0.001) GRANULARITY 4;

Thanks.