TodPunk
Joined 14 years ago
Joined 14 years ago
Unsupervised is an AI driven data analysis company. We run an unsupervised learning engine we built in-house to do a lot of analysis on customer data. The value (according to customers) is the insights we find that help them take action in their organization to move the needle on their metrics, things that matter to them and their goals.
We're hiring software engineers to work on our core compute engine and web interface. In the web app we use React and Graphql via apollo-server and TypeScript server-side, speaking to GRPC services we write in Go and Rust. In the compute engine, we're largely Python and C++. You don't need to know those coming in as long as you can learn it fast enough. On the compute side, Dask and Ray experience is a plus. You definitely need to know data structures and graphs enough to talk shop about them. We deploy on Kubernetes in a cloud native modern dev workflow, but we'll happily train what you need to know there, we've got decent tooling in place to make it very approachable and iterative.
See open positions and specific descriptions here: https://grnh.se/63f1772d2us
Unsupervised is a Data Capitalization Management company. We help organizations find actionable insights in their data so they can make their data more of a measurable asset (think extreme leverage from data science). We do this by using an unsupervised learning engine we built in-house to do a lot of analysis on their data.
We're hiring software engineers to work on our core compute engine and web interface. In the web app we use React and Graphql via Graphene (Python Graphql server). In the compute engine, we're largely Python and C++, but you don't need to know those coming in as long as you can learn it fast enough. On the compute side, Dask and Ray experience is a plus. You definitely need to know data structures and graphs enough to talk shop about them. We deploy on Kubernetes in a cloud native modern dev workflow, but we'll happily train what you need to know there, we've got decent tooling in place to make it very approachable and iterative.
See open positions and specific descriptions here: https://grnh.se/63f1772d2us
There's no rules in business but the laws of the land. Pretending that there's one definition of success is very limiting to both focus and opportunity for all of us.
This isn't exactly an HN love story, but it's working out well for us so far, and we're learning a lot. I can't say I regret it, and at the moment I know that I would regret if I had gone after funding of some sort. Right now it's just me and the customer, working out who needs what. Third legs need not apply.
Fees sucked, so I started (shameless plug) Phone Janitor[1] to give me better filtering. This allows me to check voicemail anywhere, unlike my cell line (unless I sign up for their service) and just transferred it to my own company. Turns out this helps a lot of other folks with similar privacy desires. Control is a wonderful thing.
Ultimately it's really, really crappy, but the Telecoms and the FCC don't want to be too heavy handed in this, and advertisers are finding new ways of skirting the donotcall rules. They ruined SMS in this regard[2] as well, deliberately.
[1] - https://phonejanitor.com/ [2] - https://en.wikipedia.org/wiki/SMS_spoofing
Nobody gives hell to every other OS manufacturer that will clobber our current bootloader configs. That means more than OSX, btw.
Of course, this is one reason why I'm not in politics.
- What kind of investment returns did the stock of those bankrupted yellow pages make and at what cycles?
- What does it really take to kill an old technology?
- Is there a way to compete with established dying tech companies at their own game, or is it purely by trying to advance their users to your new tech?
- What does this say about transitional companies that possibly offer both the old AND the new techs, like the yellow pages that offer online versions? Are they going to move forward or are they just delaying their deaths a bit?
- What kind of talent is needed to sustain these kinds of businesses? It's going to look very different from the talent that grows new business, but I can't deny that they're both forms of talent.
Lots of food for thought.
The reason VHS and DVD sales are better than digital is because it is still vastly easier to put a thing in a magic box hooked up to the TV and press play on the remote. Even this is frought with danger (How do I hook it up, how do I get to the right input, where exactly is the button to play the movie on this DVD menu?) This is also why you hear those complaints near constantly.
There's other factors as well (like people not feeling like they own a file, but a DVD is theirs), but again the barrier to entry for consumers on computers is getting past having to think of any of the moving parts. This is true in a lot of places, like automatic transmissions and ATMs (read up on the history of ATM interfaces, it's fascinating if you're into that).
Currently on a PC if you download a file that is a piece of media to consume, you have several important barriers for most people: - I have to put the file somewhere, and I don't understand filesystems - I have to have software that uses the file format, that works on my machine - I have to know how to use that software - I have to know how to purchase the file, whatever that is supposed to mean - I have to like how it's being displayed - I have to know how to get it from the machine I downloaded it on to my TV or whatever
There's probably others in some scenarios, but you probably see the point. A lot of things we tech users take for granted is that these things are very, very hard for most people, and they always have been (again, physical tech like cars or elevators follow this). Even giving someone a link is barely beginning to scratch the surface of this process. So this is, bar none, the reason digital is going to have problems until netflix or youtube or whatever can just play everything.
This is also, as a related aside, why a web browser and mobile apps win in breaking down those barriers. Everyone already has them and it's the "click icon to do thing" model, where icon might be a bookmark or typing a simple address, but the point is the same. No installs, no understanding of local mechanics and difference. It's also why Apple products are viewed by the consumers as more user friendly. There's very little to understand about your machine on a Macbook. You don't have to take my word for it here. Watch your average user use a mac. Or read their user interface guides.
I've written specs for studio approved DRM schemes and done pre-dvd content on in-flight entertainment systems. Please believe me when I say that it is a nightmare even if you DO know what you're doing with all the tech involved, but even if it weren't a legal problem there's a good example that is pretty damning as a foundational assumption: we still don't know how to share a file.
I mean that sincerely: try to get your average non-tech worker to send a file to your grandmother on their own. I can not overstate this enough. This does not work.
We have a lot of distance to cover before the digital content world cares about our niche whims.