I have been building my own mini version of Mattermark as a curiosity project. I'm not even interested in deal flow as much as being able to predict rising startups. In terms of being a software engineer, I look at this is a career tool.

I'm sure the bigger VCs and accelerators are building interesting things in this space but I haven't found much that's open source.

As data points, AngelList and Crunchbase are good starting places, but the AngelList API is private and the Crunchbase API is paid.