[Radim]

Employee => academia (PhD in AI) => open source (Gensim) => ML freelancing (radimrehurek.com) => ML consulting company (rare-technologies.com) => ML products (pii-tools.com, scaletext.ai).

Looking back, it's hard to imagine what it would be like to jump right into full-time products, skipping the intermediate budding steps. Is it a burden to understand the whole process, from accounting, legal, HR, management, ops, sales, support? Better to outsource them right away?

It's definitely true that all that ancillary stuff is a distraction, a (stressful) time sink. Especially when you're just starting out and clueless, like I was. Bootstrapping slows you down. On the other hand, it felt kinda natural, but took many years. Not the standard (?) SV path of rapid growth.

[alfonsodev]

How old were you when quoting as an employee and started the PhD ? Was it hard financially ? I’m asuming you had a degree already.

[Radim]

Quit in 2009, age 28. It was not that hard financially -- I moved to southern Thailand, worked on the PhD and Gensim there. Thailand is a pretty cheap country to live actually.

[abrichr]

How did you negotiate doing a PhD remotely?

[Radim]

I feel like I'm not helping you guys much, because my circumstances may have been unique, but the honest answer is:

Not much negotiation needed. My uni had what I'd call an "old school AI department", grounded in linguistics / semantic web and RDFs / expert systems, that type of research. I was more into statistics, Bayesian inference and vectors. The mutual impact was lukewarm, and completing my thesis remotely not an issue.

[zerr]

Did "PhD" pedigree help in anyway? I mean you could do your research outside academia, since your department was not helping you (in research). But I guess US PhD stipend could be considered as a good salary in Thailand.

[pbnjay]

Most programs have a year or two of coursework, then it's just research and writing. So if you're ABD ("All But Dissertation") it wouldn't be a big deal if your advisor is on board. You can do committee meetings over videochat.

I was ABD and took a job, and finished my dissertation in the evenings. I basically just had to fly back to school to defend.