The use of indexes in aggregations is not obvious. However, they will only be used if $match is at the top of the pipeline. Computed values from within the pipeline cannot use indexes. In your example, addDoc is computed from $unwind and so cannot use an index.
Maybe you could rewrite the aggregation to run against Address, where you could $match on name at the top of the pipeline and $lookup against Employee if there is a corresponding foreign key reference back to Employee.
There’s also a lot more which happens before a pipeline is executed, to help optimise the throughput. Check the MongoDB Aggregation Pipeline Optimisation doc for more detail.