Machine Learning in Cambridge, UK

In May of 2017 I relocated to Cambridge, UK to take on a new role within Amazon’s Machine Learning group. I am building a new engineering team in Cambridge that will work with a group of Machine Learning scientists guided by Neil Lawrence. We have a number of exciting projects to pursue over the next few years. In addition, there is the living-in-England thing, which is its own kind of adventure.

Sometimes it’s just about the fun

Our QA Lead took it upon himself to organize a Scooter-day for this afternoon. We hoped to get most of the team involved, but in the end we had to settle for the few people brave enough to race scooters around downtown Victoria with the die-hard motorcycle riders in the group.

The new riders were a bit nervous in the first hour, but it was great to see them relax and really get into the goofy mode that is scooter riding. Nobody crashed, nobody dropped their machine, and everybody had a great time.

So what crazy activity can we plan next?

Check out some photos from the ride – especially the last one. It’s truly epic!

Putting Intrinsic Rewards to Work

In my last post I talked about the Type-I personality. This personality trait is shared by most technical workers and especially by software developers. If you want software people to be happy, productive, and generate great ideas you need to tap into their intrinsic motivation.

So what have I done to put these ideas into action at GenoLogics?

Tech Lunch
About three weeks ago I introduced our team to Tech-Lunch. Every two weeks we book the two largest meeting rooms in the office for a full two hours. One room is for the “Tech Lunch” and the other is for the “Less-Tech Lunch”. One volunteer per room selects an hour’s worth of videos or presentations from TED, InfoQ, Google Tech Talks, or something similar and plays them for the rest of the attendees. When the videos are over, the remaining hour is used for a BoF style discussion about the topic and how it affects our profession, our jobs, and our products. We’ve had two iterations so far and discussed UI design with Don Norman, heard about 3-D spacial interfaces from TED, and learned about the Good Parts of Javascript from Douglas Crockford.

While it isn’t quite the same as attending a real conference, it’s far more convenient, less expensive, and more diverse. Oh, and the company pays for lunch.

Intrinsic motivation requires an opportunity to experience continuous learning. This is part of our solution.

Project Improvement Projects (PIPs)
Unless your software projects are vastly different from everything else out there, you have dozens of great ideas you’d like to see added to your product but you are never given the time or permission to make it happen. If this happens for too long it starts to affect your sense of independence and destroys your intrinsic motivation.

Two weeks ago we introduced PIPs to GenoLogics. For half a day every Wednesday, everyone on the development team (both coders and testers) work together on small, high-value projects and product enhancements that they think are important. Working as a team, they select one or two small, but great, ideas that have been burning in their minds and get them done. If their ideas will affect the customer experience must have them vetted by the Product Manager, but PM has no say into what items the team chooses to work. In short, developers have 10% of their week to focus on anything that they think will improve their efficiency or the experience of our customers.

We’ve only been at it for a few weeks, but we’ve already seen great things happening. Intrinsic motivation requires autonomy. This is another part of our solution.

Ground Breaking
GenoLogics has a long history with something we called “Hack Day”. This was a full day, once per quarter, when everyone at the company was encouraged to spend an entire day working on anything they found interesting. Telephones are turned off. Email is ignored. Everyone focuses.

Starting in 2010 we re-branded our day as “Ground-Breaking” in an attempt to encourage more people from outside development who might not be comfortable “hacking” for a day.

Yesterday was our second Ground-Breaking of the year and it was a huge success. People came in at 8am for a company provided breakfast and kick-off meeting at 8:30. Lunch was served at noon. The hacking, ground-breaking, and learning continued until 3:30pm when everyone met in the large board-room downstairs. For the next 90 minutes each person or team took four minutes to present the results of their one-day work marathon. We saw some truly innovative web-based UIs built on our new APIs, a custom iPad interface to our LIMS, a great tool to help automate the migration of configurations in our products, and over a dozen other terrific idea.

Type-I personalities require the chance to work on Mastery: the act of working at the very edge of your skill-set. Ground-Breaking is another part of our solution.

GenoLogics has made a great start, but there is still more we can do. We’ll keep working at it and I’ll share our experiences here.

What are you doing to help encourage Intrinsic motivation among your teams?

Type-I Personality

I just finished reading “Drive” by Daniel Pink. It has changed how I think about motivation and morale in the workplace.

In his book, Pink introduces what he calls Type-X and Type-I personalities. A Type-X person is externally motivated and responds well to things like performance bonuses, sales quotas, gold-stars on their report card, and carrots dangling on sticks. A Type-I person is intrinsically motived and responds well to things like meaning, purpose, doing-something-cool, building things people love, and making the world a better place.

In my experience, most developers are Type-I personalities.

So what does Pink’s book tell us about motivating developers?

The Not Surprising Things

  • Developers like to work on software that people actually use.
  • A development team will work harder on something that they believe will improve the world.
  • Learning new things and a chance to stretch their technical skills makes Type-I people happy.
  • Using cool technology makes the job more rewarding.
  • When Type-I people are encouraged in a Type-X manner, their work is of lower quality.
  • Poor salary can cause you to lose employees (Type-I people like money too), but a high salary is not sufficient to retain a Type-I person.

The Surprising Things

  • Focusing on Type-X things (sales this quarter, lines of code written) is not harmless. It can actually drain the intrinsic value out of an activity. In short, these things are de-motivating for Type-I people, even if the messages are constantly positive.
  • Type-X behavior is natural. Type-I behavior is learned.
  • Type-I motivation results in far more powerful and long-lasting effects than Type-X.
  • There is a right time for carrot and stick motivation, but it is pretty narrow.
  • Common motivation techniques often have the opposite effect.

One of the strong focus points we’ve had at GenoLogics lately has been on our quarterly bookings numbers. Keeping a close eye on bookings is critical for a growing company. However, by shining a bright light on our financial performance Pink argues that many Type-I employees will begin to lose sight of the intrinsic reasons they work for GenoLogics (such as the desire to help cure disease and improve the standard of health worldwide). In short, by trying to motivate people by showing how their work will help increase our numbers, we may be achieving the exact opposite: turning things into a Type-X activity and draining our Type-I employees of their desire to work.

That’s a pretty sobering thought.

If you manage people who fall into the Type-I category (nearly all knowledge workers and creative types) I urge you to give Pink’s book a read. It just may change the way you think about motivation.