Best way to Quant: ML Engineer in Tech or Regular Software Dev in Finance?

If I eventually wanted to do Quant at Two Sigma, HRT, etc., which route is the best: 1. ML Engineer in Google / Facebook / etc., or 2. Regular Devs at those firms, and then do internal transfer?

Option #1 pros: - those finance firms are hot for ML specialists, and greatly value engineers from G/FB. - Will have a lot of free time to learn about ML at work and outside of work (due to good work life balance -- maybe not at FB) - If the switch to Quant never happens, ML is the hottest specialization for the future. will still be good either way - no non compete to move to a finance firm

cons: - might not be able to be placed as Quants right away. - Cash comp or even total comp while at G/FB is lower compared to #2 (not including stock appreciation -- unless the bulls are on the run again like the past 5 years.. probably unlikely).

Option #2 pros: - maybe internal switch is easier. but need to be good at politics - Total comp is higher than #1 while waiting for the transfer (if it ever happens).

cons: - maybe internal switch is actually more difficult lol, especially if many devs are already trying to do the same. - Some firms actually value external hires more than internal hires. - If the hours are bad, might not have time to catch up with ML outside of work - non compete is punishing -- other finance firms are less likely to hire you (unless you're a superstar and they're willing to wait like 1 year). if the switch doesn't happen then worst case will need to move back to tech again..

I'm leaning towards #1, but would like to hear some opinions as well.