I would definitely say no to a job based on ethical grounds
Anna Geiduschek is a software engineer building data tools for criminal justice reform at Recidiviz
Quick profile on Anna
- Location = San Francisco, California
- 6 years since graduating from undergrad at Stanford University
- Senior software engineer at Recidiviz, past experience at Dropbox
- Anna is a developer working to collect and understand criminal justice data from state governments for influencing change in the system
- Definition: broadly speaking, think of recidivism as the situation where someone was once incarcerated, then they were out of the system, and then became incarcerated again for any reason. In the criminal justice space, recidivism usually refers to the act of being returned to some form of incarceration as a result of an action, which may or may not be criminal in action. For example, someone could have their parole revoked for using mouthwash with alcohol in it.
- Interview date = Feb 24, 2020
How did you get to where you are?
I was always a huge nerd about math and science. My dad was a mechanical engineer and he worked in the movie industry doing special effects like animatronics. As a result, I always thought engineers do really cool things like working on movies. I didn’t really know what the reality of being an engineer was, but it was always in the back of my head that I wanted to do something with science and math. I went to a school that was a bit of a hippie elementary school where they didn’t really focus on math or science. My parents could tell I enjoyed those areas and was frustrated by a lack of challenge. They put me in an online math program that I enjoyed a lot.
I took an AP computer science (CS) class in high school, which was awful, honestly. The teacher was terrible and was probably reading one chapter ahead of what he was teaching us in class. Even though I kind of liked CS as a subject, I had the sense that I could never actually do it professionally since the class was such a joke. Then I went to Stanford undergrad. When I first started, I thought I was going be a mechanical engineer because that’s what knew from my dad. But, everyone in my freshman dorm was taking CS because it’s just a class that everyone takes in undergrad. Having had exposure to CS before, with the AP class, I wanted to do that type of work again and I just took to it. It was an amazing place to learn how to code, since the CS department hires lecturers to teach rather than professors, so the teachers are people who actually want to teach instead of it being an add on requirement for them.
I had one summer where I did CS research at a university in Switzerland. Then the next summer I had a “real” internship at Palantir. My biggest concern at the time was how do I get a job that looks prestigious to put on my resume. At the time, in 2013, Palantir was one of the hot jobs to get and there was a pipeline straight from Stanford into Palantir.
I did learn a lot while I was at the Palantir internship, but I certainty came out of the internship knowing I absolutely did not want to work there.
It was a weird place to be. Now, I’m sure people know more about it. At the time, it felt like a culture of people who were extremely stereotypical in the way a bunch of young male software engineers were expected to be. For example, on the first day I was told, “Okay, don’t forget to order and expense a nerf gun.” Looking back it was dysfunctional in so many ways. I was on this team that had 8 full-time people and they put 18 interns on the team. Everyone on that full-time team became mini managers for a group of interns over that summer.
When I was getting closer to graduating from undergrad, my cousin-in-law had just started a job at Dropbox as a Product Manager and said I should check it out. I applied and got the job and started on the team building Carousel, Dropbox’s photos app which was shut down in 2015. Once Carousel dissolved, I somehow ended up as the only person at Dropbox still working on mobile photo uploads in the main Dropbox app. At first I was the whole team with basically no product management or design support, then it was a tiny group of three of us, with very limited resources. It was frustrating at times when there were conversations around killing the [camera uploads] product or keeping it alive, but I also learned very quickly to be comfortable making decisions that impact millions of people. I was appreciative to be given such a large responsibility in a relatively short amount of time, all the while surrounded by smart people who understood software best practices.
How has being a woman, in a largely male dominated field affected you?
It’s certainly something I’ve noticed and thought about since I was in school. My intro CS classes had a very even gender ratio, but there was a steep drop-off in the systems and back-end classes that I took later in undergrad. During a final exam, I could count that I was one of 6 women in a class of 200. Others [male software engineers] would recognize that I was in their class and I would not even realize they were in the class, because there were almost 200 other people that looked like them.
The whole time studying CS, and also during my time later at Dropbox, I was pretty strongly aware I was one of only a handful of women.
During the internship summer at Palantir, of the 18 interns in my immediate group, I was one of 2 women. At Dropbox I didn’t work with any [women] software engineers for much of my 4 and a half years there. There were a few on the periphery, but almost none on my direct team. That being said, I have been extremely lucky to have mostly felt supported and respected by my male colleagues.
How did you transition from Dropbox to Recidiviz (a nonprofit)? Did money matter?
With the exception of a few people, I never really felt like there were people who were “my people” in the engineering org at Dropbox. This is a bit of a broad generalization, but it felt like there was a lack of perspective of things I felt were important (social issues, politics) and the things that other people felt were important (software updates, the Apple developers conference). I had a hard time with that.
There was also an overarching existential dread where I felt nothing I was doing at Dropbox was actively bad, but I also didn’t feel like anything I was doing was particularly great. It was just fine.
After three or so years I also felt that I had learned enough at Dropbox to be an effective engineer at a smaller company with fewer resources. I started keeping my eyes out for something new and more meaningful.
I got lucky. I’m working at Recidiviz now. A close mutual friend started it as a side project that she was trying to turn into a full-time organization. I latched on and signed up for their email list and keeping tabs on what was going on. Then, once they actually got funding, I was eager to join full time immediately. I started last April (April 2019) and we did Y Combinator, a tech incubator, last summer. I was the second full-time at Recidiviz.
Money was honestly not much of a factor in making this decision:
- (1) My salary is basically the same. I’m making a lot less compared to Dropbox because I don’t get equity. But, when I look at the day-to-day cash I had in my bank account before and after the job change, it’s close to the same.
- (2) I don’t think I get that excited about having a lot of money. I’m still able to live much more comfortably in this city than the average person can and that’s enough.
I have always been a person who doesn’t want to think about money. I don’t want to think about credit card points or day trading or any of that. I just want to put it in a bank account and not think about it (I’m lucky to even have that luxury!). Sometimes at Dropbox during compensation conversations I knew I was supposed to be negotiating, but at the same time I already knew it was a wild amount they were paying me for what I was doing, given I was able to take a one hour lunch break and had top notch food in a luxury cafeteria. I think if moving to Recidiviz was 50% less pay than it actually is, I would have had to think a bit more about how I would adjust my lifestyle — would I have to move? That’s when the decisions become a little bit harder. But, right now, there’s still food on the table and money in my 401K. I don’t care much about the hypothetical SF house I’ll still never be able to afford anyways. Having more money obviously creates more options, but there’s also an exaggerated focus in our society around money.
Ultimately, it comes down to the question of would I rather enjoy and feel passionate about what I’m doing 40 to 50 hours a week, or would I rather not love what I’m doing and have more money. It just doesn’t make much sense to me to choose the latter.
What’s your role at Recidiviz?
In the US, we have the largest number of incarcerated people of any country in the world. Recidiviz’s goal is to give people data to help make decisions to ultimately end mass incarceration.
There is little to no aggregated data for the criminal justice system.
What we do now is work with state governments. State governments agree to give us their data about their prison and supervision populations and we transform it to a centralized model for them to then interpret their data. We see ourselves as a lever that will help the people in power make more informed decisions to make positive change. Our end users might just want to know how many people were admitted to prisons this month and strategies to reduce prison population sizes. They don’t have any live numbers to back up those types of decisions. They may have raw data in a location that only one person has access to or knows how to interpret.
Right now, Recidiviz is lucky in that ending mass incarceration is a rare issue with pretty widespread bipartisan support. Where the arguing comes in, across political lines, is more around the best and safest approach to fixing the system. Recidiviz can help with this in two ways — the first is to provide baselines and data to set goals and the second is to then use data insights to determine what does work. As an example, with enough data, we might be able to show states that basket weaving creative programs do little to help people succeed and stay out of prison once they leave, meanwhile working towards a GED in prison does help with reducing recidivism.
My job is still very much as a software engineer. I work more on the end of integrating with the states and turning data into something we can understand. There is another side of doing calculations and insights on top of the data and building the actual applications. I don’t do as much of the latter right now.
We’re basically an engineering-only team at Recidiviz. We have our cofounders, who are also the executive director and CTO. Then there are three full time engineers, a data scientist, and a PM. Now we’re just getting to the point where we’ve hired someone to help hold the states accountable to actually use the data and check in with them. Google has a fellowship program where they will lend a group of engineers to a project for 6 months. Before I started, Recidiviz was able to get things off the ground with that Google program. Those engineers were offboarding while I was onboarding to the company.
What are some major challenges you’ve faced at Recidiviz?
At Dropbox, everyone was in this tech world; they spoke the same language, worked at a certain fast pace, and had a high level of technological knowledge. At Recidiviz, we work with state governments where we cannot just send over a Google Doc. You would have to email a Microsoft Word file. Additionally, the expectation around how long something takes is completely different. I’ll hear that they’ll get back to us in a week, for an ask that you would imagine someone could text back about within a day.
Then there is the challenge of dealing with extremely old data systems. For one of the states we’re working with, in order to get access to data we have to remote log in to their IBM mainframe. Then in the database, all their tables are named as numbers and the columns in those tables are three letter codes. In order to decipher this, they mailed us a paper binder with the key in it. Some of this is really unglamorous compared to what people expect tech to be. At the same time, it’s a puzzle.
The data can be depressing since it’s about our prison system and those who are incarcerated, but it’s interesting since it’s real lives and real things that are happening in another state. It’s not just some abstract problem.
Another thing was that at Dropbox there was a bias towards having everything correct on a product. This was mostly because the product was in the hands of millions of people and there’s a lot more users angry if something breaks. I’m getting used to the mode where I may have to be a little bit scrappier and sloppier to make impact faster. I always have to remind myself that the faster my work gets out there, the more impact it will make.
I was actually surprised early on, at Recidiviz, that a lot of the problems I was solving were actually much more interesting than at Dropbox. At Dropbox a lot of time was spent piping data from point A to point B and winding data through the system without breaking any other code. Also, a lot of the design algorithms were already designed. Now, I get to think a lot more about interesting performance issues and schema designs, which is essentially understanding the actual system of incarceration and supervision. I think about how situations actually relate to each other. For example, if someone has multiple sentences, can you be on parole and probation at the same time? How would you model that if probation gets revoked while you’re also on parole? Then there are others days when I’m just trying to figure out what the difference between status 458101 versus 458102 means and that can be really annoying, but it’s also exciting to finally understand and crack what’s happening.
How did you get into local tech politics on Twitter?
I get a bunch of recruiting emails, as anyone who is a software engineer does, especially if you’re a woman. I rarely paid much attention to them, and mostly felt overwhelmed by how many of the companies were doing less meaningful or more destructive work compared to what I was doing at Dropbox.
I was very fascinated by tech being able to use it’s leverage as a platform, that everyone uses, to then take a stand. I knew at time that Mijente and RAICES were working on a campaign to get Amazon to drop Palantir as a customer of their AWS platform, since Palantir works directly with ICE. Now Amazon has direct contracts with ICE, which wasn’t a thing we knew for sure at that time.
Anyways, then an email came into my inbox from Amazon. Having been at Dropbox, I knew how hard it was to hire really good engineers and how much leverage engineers have over recruiters. I sent back an email to the recruiter explaining that I couldn’t join Amazon until they drop their Palantir contract and tweeted about it. I only had 3 twitter followers at the time and it kind of blew up.
What I didn’t realize was how wild it was, at the time, to say yes or no to a job based on ethical grounds.
In my mind, I would definitely say no to a job based on ethical grounds. Everyone else was freaking out about it because it was Amazon (a prestigious job to get, by most standards). That experience made me think more about how much power tech workers have over employers. Where you work, and where you choose to work, matters. Saying no to these tech companies that aren’t making the right decisions is important because the work needs to be done to show that potential employees will refuse to support a company like that.
The thing that’s crazy to me is that software engineers in this town (San Francisco) have so much power. I understand there are people who are here on a visa or have other situations where they can’t speak out. For others, if you want to speak out about something, you will still get a job. I actually still get recruiting emails from Amazon to this day. These big companies don’t know or care what you’ve said, as long as you’re going to work for them. So, if you do care enough about something they’ve done or are doing, why would you want to work for them?
What’s one tip you’d share with others on how to merge social impact with work?
I personally do not find it easy to have intense extracurriculars outside of my job. I wish there were more organizations where they paid you to do things that were of benefit to society. They are out there though. One way to find them is to just start to keep an eye out for those types of organizations and try to become involved without putting too much pressure on yourself to be super productive. You can get a sense of whether you’d like working with these people or if the work is exciting, similar to what I did by joining the email list at Recidiviz as a first step and then sitting in on a weekly status update call to see what’s going on there. I would say I have no status updates and just listen in.
Another way is to learn more about what you think you’re interested in. This could be reading a book about a specific topic or following people on twitter who are in a specific sphere. These are just areas to give you context for later, when you run across an organization that might be interesting to you.
If you could change one thing about the world overnight, what would you do?
Fundamentally, we have made it okay for resources to be distributed in an inequitable way. You can solve most problems by redistributing wealth between people and then that will redistribute power.
References
- thanks to Mathew Antony for connecting me with Anna
- Recidivism definition and measurement by the National Institute of Justice
- Dropbox camera uploads
- Dropbox — sunsetting Carousel
- Y combinator (Recividix s2019)
- Fast Company article on Recidiviz
- New York Times article on ‘Techlash’ Hitting College Campuses, featuring Anna
- Recidiviz website
