HNHacker News
TopNewBestAskShowJobs

tinkerdol

157 karma · joined May 3, 2013

submissionscomments
tinkerdol··on An Argentinian in Stuttgart: Founding a Startup in Germany
"you have to still try to do better if you want to keep it that way ;)" roughly means: Nothing is ever perfect but we should actively seek continuous improvement, as well as try hard to identify weaknesses in order to correct them. After 3 years here, I am slowly learning. Wie Rewe, jeden tag ein bisschen besser.

edit: this is why it is typically un-German to boost, and more typically German to complain -- they are just pointing out what to fix!

tinkerdol··on An Argentinian in Stuttgart: Founding a Startup in Germany
How about here? http://startup-stuttgart.de/18th-lean-startup-meetup-stuttga...
tinkerdol··on An Argentinian in Stuttgart: Founding a Startup in Germany
The new law you mention is the new "Bluecard" law: http://bluecard-germany.com/

Here is a more useful site explaining how foreigners can come and work in Germany: http://www.make-it-in-germany.com/en/

But yeah, I don't see anything about founding a company...

tinkerdol··on Disruptive Innovation vs. Harvard: Who Will Win?
>"Measuring" educational products is functionally impossible.

Why? I'd agree it would be difficult, but "impossible?" For instance, I could imagine creating tests to measure a student's competencies in things such as engineering, programming, or foreign languages pretty easily. Then once you measure a student's competence at the beginning and ending of a curriculum, you have your measurement of the educational product. Not that standardized testing is the answer, or easy for any subject... but it's an example.

>teachers keep going back to the same methods that have always been used.

Maybe if they either don't care or if they don't see a reason to change. I'm helping teach a course starting next month, "Technical English" (for non-native speakers), and am always looking for ways to improve on what I did last year. In fact, if anyone out there would like to talk to my class about your job over skype one day, I will integrate that into my curriculum right now. (You could add "guest lecturer" to your resume as a possible incentive).

tinkerdol··on Paul Graham on Building Companies for Fast Growth
>But 564 startups is enough to see patterns

This is an interesting point. Beyond happening to notice patterns yourself, have you tried running machine learning algorithms on your data to try to find other characteristics?

tinkerdol··on Defeated
Hi, sorry for the late reply!

First step is to figure out if you prefer a University or a "Hochshule" -- kind of a foreign concept. (https://en.wikipedia.org/wiki/Hochschule).

The Universities come with more prestige, are larger, offer a wider variety of programs, and are more theoretically based. Therefore if you like computer science theory, taking classes that are math-intensive and could include things like algorithm optimization or building compilers, a University is the clear choice. I have heard only awesome things about TU Munich and TU Darmstadt, although any of the big TU's are probably great.

The other option, a Hochschule, is more practice based. Although there will be a mixture of theory and practice at both types, the distinction lies in the emphasis. So if you want to spend more time going straight to writing code, a Hochschule might be the better choice. The other advantages of a Hochschule are that they can be focused on a particular topic (if you are looking for something specific) and they definitely will have smaller classes. This means the professors know you, and you go directly to them for help for which they are often freely available (contrast with a Uni -- usually either go to a TA or have to schedule an appointment with the prof a week ahead through their secretary etc.)

I went to a Hochschule and was beyond pleased. There were only 20 people in my whole program so the individual help you can get from Professors was amazing. I actually studied Geoinformatics not Computer Science directly (and therefore searched for that specialty), but I think any of the larger cities will have good Comp Sci M.Sc. possibilities at their Hochschules.

In any case you can always visit ahead of time if you can to talk to the professors and sit in on classes. If that's not possible they are also very responsive to calls and emails... they often have preferred quotas for foreign students and love to recruit them.

Hope that helps!

tinkerdol··on Defeated
May I suggest Germany? At least in regards to your first point, they have recently relaxed barriers. Right now all you need is a four-year degree in order for you to come and look for work for up to 6 months. The chances of anyone reading this finding work are extremely high because there's a shortage of skilled professionals.

http://www.make-it-in-germany.com/en/

Or, start out studying for a Master's degree first to get comfortable with the country. (Many are taught in English,and would be free. I can recommend certain programs if anyone is interested).

tinkerdol··on Machine Learning Cheat Sheet for scikit-learn
I was thinking DBSCAN or OPTICS could also be used when the number of categories is unknown, although I am no expert when it comes to understanding what sample size is necessary.
← PreviousPage 3 of 3