50% is ... way off the curve. 20% is about average in the United States, but it goes down when you exclude roles that tend to be managerial to something like 14%, so your experience seems to to track with statistical norms. Just stay in the field, people need examples and role models... building up a base of role models is a really, really hard problem because it sucks for the first wave.

This problem was identified in nursing years ago (inverse). Nursing has great opportunities for growth, pay roughly equivalent to programming and shortages in every state so severe states are starting governmental programs just to get more nurses... yet less than 7% of nurses are men due to a variety of reasons, but the major one is lack of role models. Despite identifying it and trying to offset it, male nurses are still exceptional rare because they never broke critical mass on role models.

Although, thinking about it, that mark may be for the whole company and not just its eng. I think a lot of it depends on how much of an effort is taken to make females happy and to recruit them. Or if there is a female involved in the early stages of the company. Still that 15% mark seems low. It may be worth discussing with your management about how to make the workplace better and how to recruit women better. Maybe mentorship programs or hosting 'women in tech' events I have to admit this discussion makes me feel weird. It's a bit odd to think about people based on their attributes. It's like thinking about coworkers based on their height.

I don't think there are many large companies with 50%, but they at least have a significant amount of female engineers. I thought Google was at ~40% but I can't find the reference anymore (searching just shows Google+ stats). I know 10gen (MongoDB) has at least some female engineers just because one of them helps run the local python meetup

Google? 40% female in eng? I don't think so. You wouldn't be able to find that number online since it's embarrassingly low, and they guard it jealously even internally.

In 4.5 years on 3 different projects (SRE team, dev team, different dev team), I was all alone. I should tally up all of the names of my coworkers and see just how it worked out. People came and went but it was always guys replacing guys.

Agh, I was hoping I was wrong. This is more in line with my experience.

It's a shame they don't release the numbers because it's easy for "40%" to float around the Zeitgeist, concealing a very real problem.