4 min read
Thoughts on my past (2018) and future (2019~)
I finally decided to start posting on Medium, sure took me long enough. My past blog on Blogger is still there (only 5 posts in 10 years.. hmm I guess I’m not very prolific).
Randomly decided to apply style transfer (deep learning) to a photo of myself
I have a bad habit of being generally dissatisfied with my accomplishments, but I should be thankful for the opportunities to do what I did in 2018:
- Built the data science team at SmartNews from scratch (still hiring in Tokyo and San Francisco). We’re now 5 members doing analytics (and a bit of machine learning and data engineering) across the whole company: product, content, marketing, sales, etc. All the team members are talented, friendly, and multilingual, and I’m lucky to be leading such a team. As a company, we became the biggest news app in Japan and the 3rd biggest on Android in the US (if we use certain convenient definitions).
- Became a mentor for Google Launchpad Accelerator Tokyo, which is an accelerator program for “Japanese startups focused on using AI/ML with a global mindset.” It was really fun and fulfilling to advise these innovative startups with huge potential, and I learned a lot as well from them and other mentors.
- Became an External Director (社外取締役) of MTIC, a company started by an old friend. The company helps blue collar workers (mostly from the Southeast Asia) find jobs in Japan. This is actually a huge issue for the Japanese government; in 2018, the government made big progress in terms of planning to open up and support many more blue collar foreign workers, given the demographic / labor crisis the country is in. You can hear more described by the CEO in the Disrupting Japan podcast. In the long term, we’d like to use machine learning for job matching, but we’re not quite there yet…
- Helped prepare my wife (Shirei) for Google interviews.. and she got in! She now works on scalable user support (i.e. how to keep millions of users happy) especially for the Pixel 3 phone in Japan. She has better work-life balance now too, and she can see her sister (who also works at Google) more often. By the way the Pixel 3 is the best phone I’ve ever used; it uses rad machine learning to take amazing photos. Especially the Night Sight feature is way better than anything Apple/Samsung etc. has made.
- Attended 3 cool conferences (Google Cloud Next, Nvidia GPU Technology Conference, and PyData LA), all for the first time. All 3 involved lots of talk about machine learning and how it’s becoming increasingly easy to be used practically by more people.
- Attended various fun machine learning events in Tokyo (Team AI, Machine Learning Tokyo, ML Casual Talks @ Mercari). One awesome event was the hands-on workshop on GANs (generative adversarial networks). Sometimes I can’t believe how much great education I can get for free. Big thanks to the organizers!
- Had lunch with David Ha, the head of Google Brain Tokyo. This guy is basically my hero. He went from being a managing director at Goldman to being a prolific machine learning researcher (which kind of started from a side project he had) and his work is really creative and amazing and sometimes mind-blowing. Thanks David for the advice~
- Wrote this Medium post in which every single bullet point mentioned machine learning (are you getting sick of it yet? yeah? well too bad! =P )
- Completed Zelda Breath of the Wild without any use of machine learning. (well, that’s not exactly true.. I relied on Zelda-related Google searches and recommended Youtube videos etc. which are powered by ML.) Really great game. Thanks to my wife Shirei for buying the Nintendo Switch and Zelda for me ❤
- Travelled to Kanazawa 金沢, Kawaguchi-ko 河口湖, Seoul, and Taipei (for weddings / vacation) and San Francisco and LA (for work / conference). In SF, I did fun office tours of Airbnb, Slack, Twitter, and Mercari. (thank you Naoya, Shiho, and Fabien!)
- Interesting books I read: High Growth Handbook (for mid/late-stage startups) and Developer Hegemony: the Future of Labor
- Good entertainment: Westworld season 2; Attack on Titan 進撃の巨人 season 3; Devilman Crybaby
- Read more books. I didn’t read enough books in 2018. My time spent on news articles / blog posts was maybe 10x that of books..probably not the best ratio.
- Study and apply more machine learning. I actually didn’t make as much progress as I wanted to in 2018 in terms of completing online courses, reading papers, etc. although I did still learn a fair amount at work and otherwise. In 2017, I was mostly in Finland, so I had many dark, cold, lonely nights and weekends to myself to focus on studying; in 2018 I had too many distractions. The field is progressing quickly, and I need to step it up in 2019.
- [typical new year resolution about exercising more / becoming more healthy]
- One more goal.. this is a secret for now