Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Why Trifacta is teaching humans and data to work together (gigaom.com)
34 points by cristinacordova on Oct 4, 2012 | hide | past | favorite | 10 comments


From what I gather I think that they are building some sort of visual tool for the easy munging of large data files (ie. converting semi/un structured data into some sort of structured form).

I would assume it would be some sort of GUI that captures the power and expressiveness of Perl in an easy-to-use interface.

Pretty curious on how they plan on doing that, since the problem is very fuzzy and can mean different things to different people.

(With their pedigree, I think the worst that can happen to them is an 'acq-hire'. Not a bad worst-case scenario ;)


Happy to see folks here are as excited as we are! Our founding team has put in many years of research on the kinds of problems we're now tackling commercially, including the stuff @pgbovine mentioned and more. Building beautifully useful data software is a major technical and design challenge. It's not just designing interfaces. It's not just building systems. It's not just modeling and inference. It's all three, and especially the way they interplay. Hence the name Trifacta.

If that sounds exciting to you, and you are interested in a well-funded, rational, early-stage data startup ... we're hiring. Based in SF, across the street from CalTrain. Drop us a line. jobs@trifacta.com


Our startup, Selera Labs, is working on a product that sounds similar in some respects. We leverage elements of enterprise search and machine learning to provide continuous monitoring style analysis over ERP transactional data.

The human interaction component is something we have been very focused on. We support human annotations (tagging and workflow) over machine generated results that allow us to tune analysis processing for false vs true positives. We've also devoted lots of effort into making the UX easy for business users to very quickly understand why a result (or exception) was generated.

http://dataignition.com


These guys are all seriously bad-ass at what they do. I had the honor of working with them back in grad school on http://vis.stanford.edu/wrangler/


Phil's contributions on Data Wrangler were amazing! And if you haven't seen his Burrito tool, you should check that out too ;-)


That's a very interesting problem Trifacta is trying to solve, I could have used it in my former job!

The article itself is a bit long, so I made a summary: http://api.tldr.io/tldrs/506d8f1a78b3cba85f000477


Thanks for the tldr version, the original article just felt too long.


As far as their team goes[1], they have the most prominent bigwigs in the field of Machine Learning and HCI working for / advising them. Looks like a promising startup. [1] - http://www.trifacta.com/team


I took a class with Joe Hellerstein at Berkeley. This sounds awesome!


I am really excited to be working with the team at Trifacta. They are the best team I have been around and they are going after a big problem that's near and dear to my heart.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: