Concentration with ion exchange resins, and regenerant being treated by precipitation would be an optimum solution. In your situation, it cannot be assumed as suggested above that the trivalent will be cationic in valence. Trivalent Chrome oxide can also exist as an anion, depending on pH of the solution.
Assuming you have 20 ppm in 1000M3 of water, that would take a large precipitation system. However, if 1/3 is trivalent, as per your estimate, and, we assume it is cationic, you would exhaust approximately 12 CF of strong acid cation resin, generating approximately 2.0 M3 of waste with trivalent chrome. The hexavalent will be a chromate oxyanion and can be picked up by a type I or type II strong base anion. If the pH is acidic, a weak base anion in the free base form, will also remove the chromate (as chromic acid). Assuming 2/3 of the volume is hexavalent, you will exhaust approximately 10 CF of this resin, generating another 2M3 of waste. This 4M3 waste can then easily be batch treated in a 1500 gallon tank. First step is to acidify, then add reducing agent, then adjust pH up. Procedure is well documented in EPA literature.
Hope this helps.