Best Database Choice for Geospatial Queries with Multiple Filters

Hey there! I'm developing an app that primarily uses DynamoDB for data storage, all managed via Lambda functions. I've hit a bit of a snag with one particular use case: I need to perform proximity searches for items by their latitude and longitude, letting users filter within a specific radius (like 10 km) alongside other criteria (like creation date, object type, or target age). Given that DynamoDB operates best with a single partition/sort key configuration, this is proving to be a challenge.

I've looked into using a geohash as a sort key, but it's a balancing act:
- A shorter geohash leads to fewer partitions to query but requires filtering out lots of results that fall outside the radius.
- A longer geohash offers better accuracy but means querying several adjacent hash keys for the desired area.

I also thought about creating a separate 'query table' in a different database that holds the attributes I need (latitude, longitude, etc.) along with the DynamoDB IDs. I could query this table without the constraints of DynamoDB, and then use BatchGetItem to retrieve the complete records from DynamoDB.

So, my main question is: what's the most cost-effective database approach to handle geospatial queries with multiple filters? Should I consider a specific database for this scenario, or can I make DynamoDB work for this use case despite its limitations? Thanks for any guidance!