Here is what I would do:
1) charge all cells to 3.65 until current drops to C/20. You can do this cell by cell or in a big parallel group if you have good temporary connetions. Let cells rest for 24 hours. Measure voltage again. Good cells will still be 3.45 to 3.5 while
marginal cells will have dropped to 3.3v. Set aside any cells that fail this test.
2) pick random cells and put into groups of 11 cells in parallel (temporary but good connections).
3) discharge each 11 cell group to cutoff voltage (discharge rate not critical) and record capacity.
4) Take least capacity group apart and charge/discharge each cell to find least capacity cells
5) Take most capacity group apart and charge/discharge each cell to find most capacity cells
6) swap around cells from most/least groups to even out their group capacity
7) repeat 4,5,6 until the variance between most and least capacity of all the 11 cell groups is small enough.
Build up pack with cell groups and you are done.