The direct reaction with HCN seems to be how the mandelic acid was synthesized in the 19th century. If you check how the reaction with sodium bisulfite is actually performed (Organic Syntheses, Coll. Vol. 1, p.336 (1941); Vol. 6, p.58 (1926).) you'll see that this is not more complicated than the direct HCN route as those first two steps are performed in one reaction step.
From this preparation, several preparations via the HCN route are cited:
Naquet and Louguinine, Ann. 139, 299 (1866); Wallach, Ann. 193, 38
(1878); Spiegel, Ber. 14, 239 (1881); Ultee, Rec. trav. chim. 28, 254
(1909); Rule, J. Chem. Soc. 113, 12 (1918).
I looked at the first page of the oldest of those papers from Naquet and Louguinine from 1866 and it states:
Bekanntlich bildet sich diese Säure bei der wechselseitigen Einwirkung des Benzoesäure-Aldehyds und der Cyanwasserstoffsäure bei Anwesenheit von verdünnter Chlorwasserstoffsäure gemäß der Gleichung:
$$\ce{C7H6O + CNH + 2 H2O + HCl -> NH4Cl + C8H8O3}$$
Translated, this means that you can synthesize mandelic acid from benzaldehyde, HCN and HCl.
I'm assuming you're not planning to actually synthesize it using this route yourself, I would guess that avoiding the use of HCN directly was one of the motivations behind the synthesis route using sodium bisulfite.
$%edit$