Well first I would say that the product you're synthesizing is commercially available, though it is a bit pricey. Depending how much you need and where you are located, it might be possible to simply by the reagent as it's not prohibitively expensive (£211 for 5g or £68 for 1g on Fluorochem).

So there is literature precedent for reducing the aldehyde to alcohol and converting to the primary alcohol and then converting to the alkyl bromide using other brominating reagents such as using PBr3 (72% yield) or NBS (47% yield) [see ref. 1 and 2, respectively]. So perhaps you could consider that, especially if you've already isolated your primary alcohol intermediate. However, if you want to avoid this route all together, the next thing that comes to mind is the Stephen Reduction.

For the Stephen reduction of a a nitrile to aldehyde (R-CN --> RCOH) yields are good to excellent. I've found the following example which is quantitative [ref 3].

Here is the procedure for that (see picture for reference):

"To a solution of α-bromomethyl-tolunitril (15.3 mmol) in toluene (30 mL) was added DIBAL-H in THF (1.0 M, 1.4 eq.) at 0°C within 30 minutes. After stirring at 0°C for 2h, the reaction mixture was poured into a mixture of 40 mL of methylene chloride and 100 mL of 10% HCl. The resulting mixture was stirred for 1h and the organic layer was washed with water and then brine, and the aqueous was extracted with methylene chloride twice. After dried over Na2SO4, filtered, and concentrated, the obtained semi-oily product was stored in a refrigerator to give 3-(bromomethyl)benzaldehyde as a white crystalline solid in quantitative yield." [ref 3]

Here is an additional reference which might be easier to track down than a patent [ref. 4]:

" 1. 1. 1. 3-(bromomethyl)benzaldehyde (6): Commercially available alpha-bromo-m-toluinitrile 5 (5 g, 25.51 mmol) was dissolved in toluene and cooled to 0 ºC under an argon atmosphere. Diisobutylaluminum hydride (35.96 ml, 35.96 mmol, 1 M/Hexanes) was added dropwise to the solution, upon which the white opaque solution became clear and yellow. After 2 hrs, 66 ml of CHCl3 was added to the solution, followed by 200 ml 10% HCl. After stirring for an additional hour, the organic layer was separated and washed with distilled water, then dried with Na 2 SO 4 and concentrated to yield a clear oil, which crystallized overnight. The white crystals were filtered and washed with ice-cold hexanes, then allowed to dry to yield the desired compound in 94% yield (4.77 g). Spectroscopic data was in agreement with literature values. 2 1 H NMR (300 MHz, CDCl 3 ) d 10.03 (s, 1H), 7.91 (s, 1H), 7.83 (d, J = 7.5 Hz, 1H), 7.68 (d, J = 7.5 Hz, 1H), 7.54 (t, J = 7.5 Hz, 1H), 4.55 (s, 2H). 13 C NMR (75 MHz, CDCl 3 ) d 191.6, 138.9, 136.8, 134.9, 129.7, 32.1. " [ref. 4]

There are a few other ways to arrive at your product (e.g., Wohl-Ziegler Reaction) but I think the reduction of the nitrile to the aldehyde makes the most since, and the starting reagents are relatively inexpensive.

[1] Hao, Xin-Qi et al; Journal of Organometallic Chemistry, 695(1), 82-89; 2009

Article Unsymmetrical, oxazolinyl-containing achiral and chiral NCN ...

[2] Haotian Li, et al..; Molecules, 21(7), 793/1-793/8; 2016

Article Synthesis and Evaluation of Novel α-Aminoamides Containing a...

[3] Preparation of diphenylpentadienone derivatives for use as antitumor agents;

Shih, Charles C-Y. et al; From PCT Int. Appl., 2012024282, Feb 2012 https://patentscope.wipo.int/search/en/detail.jsf;jsessionid=33BCD18F94AA5B7A1108DC7DB0B2E782.wapp1nB?docId=WO2012024282&tab=PCTDESCRIPTION&queryString=%28%28&recNum=123&maxRec=1000

[4] Tuley, Alfred et al, Chemical Communications (Cambridge, United Kingdom), 50(56), 7424-7426; 2014

https://pubs.rsc.org/en/content/articlehtml/2014/cc/c4cc02000f

Article A Genetically Encoded Aldehyde for Rapid Protein Labelling

Long answer short:

"1. 1. 1. 3-(bromomethyl)benzaldehyde (6): Commercially available alpha-bromo-m-toluinitrile 5 (5 g, 25.51 mmol) was dissolved in toluene and cooled to 0 ºC under an argon atmosphere. Diisobutylaluminum hydride (35.96 ml, 35.96 mmol, 1 M/Hexanes) was added dropwise to the solution, upon which the white opaque solution became clear and yellow. After 2 hrs, 66 ml of CHCl3 was added to the solution, followed by 200 ml 10% HCl. After stirring for an additional hour, the organic layer was separated and washed with distilled water, then dried with Na 2 SO 4 and concentrated to yield a clear oil, which crystallized overnight. The white crystals were filtered and washed with ice-cold hexanes, then allowed to dry to yield the desired compound in 94% yield (4.77 g). Spectroscopic data was in agreement with literature values. 2 1 H NMR (300 MHz, CDCl 3 ) d 10.03 (s, 1H), 7.91 (s, 1H), 7.83 (d, J = 7.5 Hz, 1H), 7.68 (d, J = 7.5 Hz, 1H), 7.54 (t, J = 7.5 Hz, 1H), 4.55 (s, 2H). 13 C NMR (75 MHz, CDCl 3 ) d 191.6, 138.9, 136.8, 134.9, 129.7, 32.1. "

Tuley, Alfred et al, Chemical Communications (Cambridge, United Kingdom), 50(56), 7424-7426; 2014

I think it is better to start your synthesis from m-tolunitrile. First of all, you convert methyl to methylenebromide with NBS/AIBN and then nitrile to aldehyde with dibal-H. Here, I have attached the reference to make your job easy.