Podcast.__init__: Building And Growing A Business with Python

CTO Christine Spang sits down with Tobias Macey, host of Podcast.__init__

As we’ve talked about before on the engineering blog, we’re huge fans of Python and it’s a big part of how we do what we do here at Nylas. So when fellow fans reach out to chat with us about how we use Python and how Nylas is built, we’re happy to take the time to chat. A few weeks ago our CTO Christine Spang sat down with Tobias Macey, host of Podcast.__init__ and DevOps Engineering Manager at MIT. They talked about how we use Python to build Nylas, what her road to CTO has looked like, and how we encourage diversity and inclusion at Nylas.

Here are a few of our favorite quotes from the episode:

Regarding why people keep saying email will die, but it never does:

By nature, email is distributed. At this point they're millions of servers out there that help people connect to the email network, and if you had your email hosted at any one provider, you can talk to anyone at any other email provider. I think that this property has been a key reason why email has become so sticky and such a key facet of how modern business communication works. There's no other system out there that essentially is an open network that doesn't require everyone to be on the same system. All of these new systems that people have been touting as email killers throughout the years lose this key property, and because there's so much infrastructure out there that keeps the email network going it's hard to replace that.

Why diversity & inclusion is so important to Christine, and how she’s translated that to her work at Nylas: 

Diversity and inclusion is incredibly important to me personally. It was important to me as part of starting a new organization from scratch that we really do our best to create an environment where all sorts of different people can be successful, not just the status quo of what's worked in the past. There are a few things that we've done to work on that:
- Having a female founder helps attract different sorts of people to the company, especially because I am highly technical.
- One thing I think that we got right from the get-go was creating a core team that really also values diversity and inclusion. It's really important that the team thinks that diversity and inclusion are important things to work on otherwise you're fighting an uphill battle. Once you do have that in place it's much easier to grow the team because your team who is on your interview panels will help find people who also buy into this overall value. 
- We also work with some organizations that are like training new folks from different backgrounds and getting them into the tech industry: Hackbright, and code2040
- I try to spend more of my time recruiting and doing outreach to women and underrepresented minorities, just because sometimes it can be harder to find people who are not the status quo, so you have to invest the time. I'm really excited to connect with more folks from different backgrounds.

On company culture:

It's really important to have a culture that is great, but it's also important to communicate that culture.

How Nylas's product has made customer's work-lives easier:

Because it means that people have to switch contexts less often, they're more effective at their jobs; they have to do less drudgery and manual data entry. They couldn't do that before because it was just really hard to connect two email mailboxes. 

Advice for founders:

I highly recommend to anyone who's thinking about starting a company to get some management experience beforehand if you can, because it's a bit of a stressful experience to be flying by the seat of your pants and trying to do the best for the company and also trying to learn all these people management skills at the same time while feeling you're not competent anymore because you're doing something completely different from what you've been doing for a number of years and feel you're actually good at.

What does the future of email hold:

I see in the future, in a five to 10 year time span, these types of more modern features being built directly into email providers so that you don't need to have these intermediate steps of, "We need to take all this data and translate it into a totally different form so that we can use it for something useful."

Second is, built-in abilities to connect parts of your mailbox to other applications. It's essentially all or nothing right now, you can either give some app like full access to your mailbox or not at all, which is not a great place to be for security. It's because when email was designed, no one thought that you were going to want to connect parts of your mailbox to 20 different applications. Email was for sending communication person to person, but email has changed a lot in the way that people are using it. When email was first designed your mailbox might have had a limit of like 10 megabytes or something like that or less way back in the day. Now, there's so much wealth of information in there that I think that the data is going to be key going forward as to what email turns into.

On the user experience side, we don't really know what that will look like. What we're excited to do as a platform is to enable people to experiment and try many different things, and then build the tools that make it so that many people can build those things easier and faster because everyone needs the same things at the end of the day. Because there's so much interesting data here and it's all very unstructured like machine learning in AI technologies will become important at some point because you essentially need to pull the signal out of the noise. We don't know exactly how that will look and excited about working with folks who want to figure it out.

Listen to the full episode here or read the transcript below.


Transcript

[00:00:13] Tobias: Hello, and welcome to Podcast.init, the Podcast About Python And The People Who Make It Great. Your host as usual is Tobias Macy and today I am interviewing Christine Spang about Nylas: Into The Modern Era Of Email. So, Christine, could you start by introducing yourself?

[00:01:47] Christine: Sure. My name is Christine and I am the co-founder and CTO of Nylas. A little bit about myself before that, I went to college at MIT. I worked a little bit at a kernel startup, called Case Place, which was started by some friends of mine, we met at the MIT computer club. Then about four and a half years ago, I started this company called, Nylas.

[00:02:09] Tobias: Do you remember how you were first got introduced to Python?

[00:02:11] Christine: I basically got into computing and programming in high school through working with free and open source software. So I started contributing to Debian Linux when I was in high school. And I started to teach myself Python for the first time as a part of wanting to contribute to Debian. I wanted to work on some package that was written in Python, so I found some online tutorials and started teaching myself Python that way. And then at MIT, I was the class of 2010, and they were transitioning their computer science curriculum at that time. I was the in the in-between year where they had a new curriculum, where the introductory courses were in Python and an old curriculum where all of the introductory courses were in Lisp. So I took a little bit of both, so I got a chance to use Python a bunch in college as well.

[00:03:14] Tobias: You mentioned that a few years ago you helped to found the Nylas company. So I was wondering if you can explain a bit about what it is that Nylas does and some of the history of getting it off the ground.

[00:03:26] Christine: Nylas is a modern rest API that makes it really easy for developers to plug email, contacts, and calendar into any application. So the idea for the company came from my co-founder, who is a friend of mine from college — he was trying to build new products and experiences that worked with email and as a part of his undergraduate thesis. He basically found that he spent several months trying to pull the data out and display it, because all of that kind of technology to integrate with email was very complicated.

So the thesis was that, one of the reasons that people haven't been building a lot of new things that have been really taking off, and making it easier to work with email is that because it is so difficult to develop with. What we took out of that was basically that we need to make it easier to work with email, so that people could be able to experiment and innovate with the ways you are able to use it. So that's how we ended up building this server infrastructure and API that abstracts away 50 years of email history and makes it so that you can work with email using the modern developer tools that people expect these days.

[00:04:52] Tobias: Email as a protocol and a means of communication has been around for a number of years and its death has been predicted time and again, but it continues to maintain relevance and stay as one of the primary means of getting in touch and maintaining lines of communication. I'm wondering if you have any theories as to what has made it so resilient in the face of so many different new protocols and means of discussion that have been developed in recent years.

[00:05:19] Christine: I think there's a few things that contribute to this. One is that by nature email is distributed. At this point they're millions of servers out there that help people connect to the email network, and if you had your email hosted at any one provider, you can talk to anyone at any other email provider. I think that this property has been a key reason why email has become so sticky and such a key facet of how modern business communication works. There's no other system out there that essentially is an open network that doesn't require everyone to be on the same system. All of these new systems that people have been kind of touting as email killers throughout the years lose this key property of email, and because there's so much infrastructure out there that keeps the email network going it's hard to replace that.

[00:06:21] Tobias: Also, a lot of the systems that are trying to unseat email as the main means of communication are largely more synchronous in nature, whereas the async aspect of email makes it more beneficial to business communications and more thoughtful discourse because you are able to take the time to write the message that you want to write and ensure that you don't have any grammatical errors or make sure that you're including all the different talking points that you want. Whereas things like IRC or slack encourage a much more rapid-fire communication, which is great when you're doing real-time interaction with people but when you want to have a more thoughtful and reasoned discourse email is a much better method of doing that because of that lack of urgency.

[00:07:04] Christine: I'm with you on that.

[00:07:05] Tobias: One of the early product offerings that you built at Nylas was the N1 mail client, which was very popular at the time that you first launched it and a little while after that-- I don't remember the exact timeline, but you decided to actually sunset that project, which is now living on in a couple of opensource iterations, but I'm wondering if you could share some of the lessons that you learned from the process of building that client and integrating it with your platform, and how that has informed your current focus and where you're spending your energies right now.

[00:07:42] Christine: I think that N1 was a really useful product for us really stress testing our API in the early days, and helped to kind of have a super heavyweight client that basically did all of the things in order to really make sure that our API worked for any sort of client that you might want to build. For that reason, I think it was totally worthwhile for us to have built it. We had that thesis in that day that-- I mean we were experimenting on what an email platform for the future might look like. The first thing we built was this data API, and that was one iteration of what the future of email as a platform might look like.

The second thing we built was this email client, which was an experiment in, "Hey, maybe people might want to instead of having to build entire applications, build like really powerful plugins into an email client that is built for extensibility?" Our experience was that-- and I think it's still possible that this kind of like vision might be a thing that could work someday, but it didn't work in this iteration for a number of reasons.

One is just like having a split focus as a company where we’re a small team or startup and having two major products it was really difficult to support from an engineering standpoint. It definitely caused a lot of kind of tension internally as to where we're spending our resources because N1 is essentially required the API to work, and yet it also required a very different skill set for folks to be kind of successful at working on it. It is hard for people to work on it on both parts at the same time. Also just hard for us to resource each project appropriately with a small team. That was one thing.

One another thing was basically that we found it very difficult, honestly, to make enough money to make up for the costs of N1 just through selling the email client. We essentially had this like longer-term vision where we were going to like plugin packages on top of N1. Those would be the products that people would be using at the end, but when you think about it and you're also already dealing with a constrained team, then you have three things.

One is an API infrastructure, which in itself is like pretty complex. Second is like the base email client, which is an extensible system, and the third is like these plugin packages. So that's a lot of stuff to support. Honestly, we just didn't really have the size of team and time necessary to build that successful business even if it had looked promising as to how we could charge for it. For base email clients, people just don't expect to pay for them these days, so it's very hard to make money just selling an email client.

[00:10:50] Tobias: The business of Nylas, looking at the About pages and the fact that you've open sourced to your employee handbook and some of the blog posts that you've written has a very strong focus on encouraging diversity and a very open nature of the company. I’m wondering if you can speak a bit to some of the ways that you encourage that type of environment and how it manifests at work and some of the challenges that you've had to overcome to ensure that that aspect of the company is continued as you add new people and the culture shifts with the inclusion of new people, and the different dynamics that they bring to the table.

[00:11:30] Christine: This is super important. For me personally, diversity inclusion is incredibly important. I came to it because I am a woman in tech and my experience throughout the years has been generally very positive in the industry, but I know a lot of people and I also know that I've been really lucky and that being early career you also don't see some of that like roadblocks that people tend to see later on, where they start getting promoted. It was important to me as part of starting a new organization from scratch that we really do our best to create an environment where all sorts of different people can be successful, not just the status quo of what's worked in the past.

There's a few things that we've done to work on that. One is just like having a female founder helps attract different sorts of people to the company, especially because I am highly technical. Few things that we do to like propagate the culture. One thing I think that we got right from the get-go was creating a core team that really also values diversity and inclusion. It's important to have buy-in from the team because these sorts of things take actual work. If people aren't excited about working with them and think that they're important then they're just not actually going to work on it. It also takes a lot of work on oneself personally examining one's own biases and doing things to address them. It's really important that the team thinks that diversity and inclusion are important things to work on otherwise you're fighting an uphill battle. Once you do have that in place it's much easier to grow the team because your team who is on your interview panels will help find people who also buy into this overall value. We try to ask everyone who goes through our interview circuits what they think about diversity inclusion and make sure that they buy into it. We also work with some organizations that are like training new folks from different backgrounds and getting them into the tech industry. One organization we've been working with is called Hackbright, which is a coding school for women. Another organization that we just recently started working with RVP of engineering mentored for them last summer and this summer we have an intern .joining our team is this organization called code2040 which is helping Black and Latinx college students.

Those organizations are really great. Trying to think of what else that we do specifically. I try to spend more of my time recruiting and doing outreach to women and underrepresented minorities, just because sometimes it can be harder to find people who are not the status quo, so you have to invest the time. I'm really excited to connect with more folks from different backgrounds. So I'm very happy to spend the time.

[00:14:50] Tobias: One of the things that can serve as either a deterrent or an attractor for people who don't necessarily fit the standard stereotype of a developer is the way that the job descriptions are written, because if there is too much of the alphabet soup or a laundry list of requirements then it can often discourage people who might not feel that they're immediately qualified for the role. Whereas with a little bit of coaching, they could very easily grow into it. I don't know if you put any special effort into the way that you write your job requirements, or your job postings when you are trying to hire for new positions?

[00:15:29] Christine: Yes, that's definitely a factor. We use this tool called Textio to look at our job descriptions. We've also spent a fair amount of effort just like writing down and describing our culture, which I think by its nature appeals to a wide variety of folks, particularly because we're a very collaborative team. It's not so much a place where folks go off and work on their projects by themselves, people are always working together. We heavily value strong communication skills.

One thing that we did was we worked with this site called Key Values, which started by another MIT grad, Lynne Tye. She really helped us work through the process of communicating and figuring out what exactly is important to us. The site's really cool. It allows you to basically browse for different companies by values. We've we found a number of women who have reached out to us just because of our profile on that site. I think that one, it's really important to have a culture that is great and then it's also important to communicate that culture.

[00:16:43] Tobias: Now digging further into the technical aspects of what you're doing at Nylas. I don't know if you can provide an overview of the way that your platform is architected, and some of the ways that Python is used within that architecture.

[00:16:58] Christine: There are a couple of major components of our system. At a high level, our server platform is essentially an email client implementation that will connect to any email provider out there. It downloads and caches folks’ email data and makes that available via your front end API. What that basically means is that, we have a bunch of API servers, they're stateless they're running application servers, all the application servers are in Python, and developers can set up applications in which their users will get bounced to our servers and go through an overflow and connect to their mailboxes, to our system.

When we get a new mailbox connected to the system, we have a separate pool of machines called the sync fleet, which in the background starts basically kind of downloading and cache of all of the email data that is in that mailbox, putting it in our data store. Our data store is basically a fleet of horizontally charted MySQL machines. They're kind of standard and set up with primary replica pairs. Each mailbox is tied to a single cluster. The SQL machines download all the data. They go through this process is basically what we call initial sync, where they have to download the backlog first. Then we maintain basically persistent connections to the mailbox provider to keep that data store continually up to date so that the data available via our API is always the latest data that's available via the email provider itself.

[00:18:41] Tobias: There are a lot of different elements of an email message that you can pull out for being able to query across. I don’t know what are some of the most used and most useful portions of the email message that people are generally accessing via the API?

[00:18:58] Christine: When we designed the API in the beginning, we wanted to really simplify the process of actually accessing the email data. To do that we basically had to decide what was important for folks to be able to easily access. If you are accessing email via the traditional protocols, there's a lot of encodings and formatting that you have to deal with in order to just drill down to an email body or to get an attachment. Our basic thesis was there's a bunch of headers that are important, the email body is important, a list of attachments is important, and everything else is secondary. We make it really easy to access and filter emails based on their recipients to, from, cc, bcc, the subject. I think there's a few other headers that we make it possible to filter on. We make email bodies available very easily via our API and we also allow you to query for attachments. If whatever you're building needs something else, some custom headers, stuff like that, we make the raw emails available via the API as well. We find that the majority of our customers find the basic representation sufficient. I think that's actually one way in which we are pretty successful in the initial design of the API, is getting it pretty right what were the most important things that people need for most things that they want to build.

[00:20:40] Tobias: You mentioned that the primary focus of Nylas is as a means of synchronizing user's entire mailbox and then exposing the data contained therein by an API. I'm wondering what are some of the typical use cases that enables and some of the ways that people have generally been using that Nylas API?

[00:21:04] Christine: The biggest, the most typical way that people use our API is for building various sorts of vertical-specific CRMs. If you ever heard the phrase, there's an app for that, what we found is that there's also a CRM for that. There are all sorts of interesting and unexpected CRMs that we found and they all need these basic features of email, and calendaring, and address books. The way that they do that is through the Nylas' platform.

For example, there's people who have built real estate CRM's on us, there's people who have built hiring applications, there's people who have built support tools, there's people who have built automotive CRMs, applications to manage salon business. It turns out that communication is a core human need and all of these tools that are specific for managing some sort of business need these features. You know how back in the day web frameworks enabled people to build web applications for doing just about anything?

There's web applications for managing climbing gyms these days and these frameworks made it so easy to develop web apps that people could do that. One of the things that we want to see happen as a result of Nylas is all of these applications basically grow these embedded communications and scheduling tools. Because it means that people have to switch contexts less, they're more effective at their jobs, they have to do less drudgery and manual data entry. They couldn't do that before because it was just really hard to connect two email mailboxes.

[00:22:54] Tobias: For somebody who is using Nylas to create the CRMs and various communications tools, is the user-facing portion generally something on the order of a contact form where somebody will land on the page, fill out the form to request some information, or sign up for some sort of mailing list, et cetera, and then the Nylas' CPI will then take that converted into an email him and then put it into somebody's mailbox? Or is it more that the user who is coming on to that site will link to their mailbox so that aspects of their email can be accessed by the person who is using the CRM?

[00:23:36] Christine: It's more the latter. Basically, these applications will have in their settings or set up, a step where folks connect their user mailboxes to that application. Then, that connection seamlessly powers some part of the application where, for example, you might have a page in your hiring app where you can see all of the previous communication the people at your company have made with a candidate and that's powered by Nylas. There might be an in-line feature for sending out email campaigns using templates or doing some kind of sales automation, and those features would be powered by Nylas. It's mostly that there's two steps; one, connect your mailbox, two, basically use all of these basic features that people need.

[00:24:31] Tobias: I've been actually on the market for CRMs myself. One of the ways that I've seen for being able to track those communications is to add a CC or BCC for a particular magic email address that will copy that communication chain into the CRM for being able to track the various steps of moving somebody through, for instance, a sales funnel Nylas. With Nylas is the idea that you don't need to use that specific email account and it will just intuit based on the person's status as a contact in the database that any emails that happen to have their address that neither the to or the from field will automatically be allocated to that particular communication channel.

[00:25:16] Christine: Exactly. We just make it more seamless, when you're developing an application you don't have to create this system that requires an email account on a mail server, have to instruct everyone to set up their mail clients to BCC properly, you don't have to deal with people forgetting the BCC. We just pull the data straight from the mailbox.

[00:25:37] Tobias: For doing both email campaigns, a lot of times you'll use some form of SMTP provider, whether it's SendGrid as just pure data channel or something like Mailchimp or being able to craft the emails and then manage to the sending. Does Nylas serve that sort of use case as well of being that SMTP channel for being able to fan out the delivery so that you're not, for instance going against the Terms of Service of Gmail and their email limits that they might have?

[00:26:12] Christine: You should think of Nylas as being a step further forward in an outreach pipeline than doing this bulk sending. We don't actually provide the functionality that these transactional email providers like Sendgrid, Mailgun, those folks provide. If you're using those kinds of services, you typically use Nylas as step two. So, do a bunch of bulk outreach then whoever is engaging with those, you might start to use Nylas as part of follow up.

We because we're using the mailboxes themselves, don't support sending out mass email campaigns. For folks that are further along in the pipeline, we provide a lot of features that basically you can't get out of these transactional email providers. One, our deliverability is much higher because we're sending out emails through the actual email account. Two, the engagement is also higher because it's sending emails just as yourself. It's just generally used for a different type of interaction where you're typically reaching out to fewer people. Because it's sent through the actual email accounts, we actually can't send thousands of emails because all of these email providers have sending limits for each day, and that's to reduce spam and abuse, and we respect those limits.

[00:27:48] Tobias: What are some of the most interesting or innovative uses of Nylas that you have seen built into some of these applications?

[00:27:56] Christine: Some of the more cool and forward thinking things that people are doing is not only accessing the email data itself, but also getting more insight from the connections between the data and stuff like that. For example, some companies are building things where you're trying to find the right person at the company to get you an introduction to someone at some other company. We have all of that connection data in our database and you can use it to essentially find out who is talking to who and basically get actually more information and insight out of the connection data itself.

On a little bit more weirder note, in the early days of Nylas we had one person who actually was at a research lab who use our API to print out all of the emails that were sent to his research lab mailing list on receipt paper and apparently the role was like a thousand feet long or something like that. It was pretty crazy. The Nylas API is suitable for everything from serious business to crazy hacks.

[00:29:15] Tobias: Given the fact of the you're dealing with people's personal or work email accounts and their inboxes and there's potentially a lot of private or sensitive data that's contained in that, I imagine that you have to have a lot of security checks and security protocols in place on the infrastructure that you're using to manage that data. I don't know if you can talk to some of the controls that you have to ensure that there aren't any data breaches or any access leaks of the data for people who shouldn't be able to view any particular aspect of those communications?

[00:29:52] Christine: Obviously, security is a really, really important part of any email platform and we take it super, super seriously. Some things that we are doing to ensure that user status stays private are: we're working on complying with all of the various different data privacy laws. There's GDPR, which is the new European Data Privacy Protection law, which goes into effect in May. We are classified as a data processor according to that law, which means we have a slightly less stringent requirements than people who are data controllers. We are well on track to providing the abilities to export people's private data and also to ensure the right to be forgotten. If anyone wants all of their data to be cleared out of our systems, they can just send a request in and we'll make sure that that is deleted within the allotted time period.

Beyond that, we're also going through the enterprise compliance called SOC 2. Which we don't have a timeline for specifically completing that, but we're in the middle of the process and it should happen sometime in the next some number of months. We also have folks, engineers, on our team who work on security and ensure things like making sure all of our external points of presence are patched and up-to–date, making sure that all onboarded employees go through security training, that folks are using encrypted hard drives, that we use the principle of least privilege and folks don't have access to the production databases if they don't need to, that data is encrypted at rest.

We encrypt all of the mail message data that's in our databases. We need to keep some metadata and headers unencrypted because it needs to be queryable. Generally, it's an ongoing process that we're continually trying to improve and already following the best industry, best practices and looking to do as good a job as is possible there in terms of keeping people's data private.

[00:32:21] Tobias: What are some of the biggest challenges that you are facing either currently or in the past or anticipate in the future whether from a technical level or from the business perspective?

[00:32:34] Christine: On a technical level, email inherently these days involves a lot of data just in terms of sheer volume. Our biggest scaling bottleneck from the beginning has always been the size of that data, that's why we had to charter databases within the first year and looking forward with regards to our application architecture, that's one of the biggest constraints. As we're looking forward to adding the next million mailboxes, we know that we have to architect our systems so that data will scale in that, so all of our services will keep operating in the face of all of that data.

We also see the next phase of product development beyond just providing the raw access to data, is actually providing tools that help people extract this extra insight and information out of the data that belongs to them. That's creating data processing pipelines, doing some basic machine learning sentiment analysis, type classification, extracting contact information from signatures. There's all sorts of things that are challenging about building those sorts of data pipelines at scale and providing a reliable service.

One thing that's tricky and interesting about running a platform, which makes it super, super fun to work on as an engineer is just the need to be able to look at what any particular users experiences. That user is a developer, they have an application that's running against us, they make some set of API calls. We need to have the instrumentation and introspection capabilities to actually verify that our platform is working as intended for any given application, which might be doing different things.

On the business side, we're looking at selling to larger companies going forward.

That brings its own additional compliance and also sales challenges. So we were working on scaling our sales team.

[00:34:54] Tobias: On a more personal level, you originally started the company, I'm assuming, as a sole contributor. Now that you are growing and adding more people, your role as the CTO I imagine is causing you to have to spend less time digging into the code and working on the technical details and to dealing more with the people management aspects. I'm wondering if you can just talk a bit about how you have found that transition.

[00:35:21] Christine: You're totally correct that, in the beginning of the company, I was an individual contributor. I contributed large parts to our IMAP sync engine and the other parts of our infrastructure API. Throughout the last four and a half years I have worn a lot of different hats as is the nature of any startup. There was a point when the entire engineering team also reported to me. I was a full time manager at one point and it's definitely been, I would say up and down.

I highly recommend to anyone who's thinking about starting a company to get some management experience beforehand if you can, because it's a bit of a stressful experience to be flying by the seat of your pants and trying to do the best for the company and also trying to learn all these people management skills at the same time while feeling you're not competent anymore because you're doing something completely different from what you've been doing for a number of years and feel you're actually good at. At this point personally, I feel I'm over the hump and it's starting to get easier.

I have five people who report to me, some of those are individual contributors, we also have a VP of engineering who's been helping us scale the management for the engineering team and that's been super, super helpful. I'm also spending more of my time these days on external facing things like working with our investors, doing a lot of recruiting to grow the team and just building relationships with folks outside the company since that's instrumental for our success going forward. There's definitely challenges to that. I am definitely an introvert by nature and I've had to figure out how to manage my time so that I can make sure I'm able to recharge my batteries.

I'm also really excited at, if you think about it, a system where the components are made up of people is also a system. There's elements of the type of thinking that you do as an individual contributor, solving problems that you can bring to building the organization itself. I'm really excited about building the machine that builds the machine, if that makes sense. It's been a bit of a journey to get there for sure, there were times when I just wanted to go back and code.

[00:37:48] Tobias: Yes, it's definitely challenging transition to make when you're so used to going so deep in the software stack and trying to build up that context and maintain the overall architecture and how everything's playing together and then being so deterministic and then having to move to trying to translate that understanding to people who are very much non-deterministic and have so many different variables that you have to account for and how the overall team is functioning and making sure that they have what they need to be able to learn and grow successfully.

It's a challenging transition to make and I respect anybody who is able to make that transition and be successful in that role. So, congratulations on that.

Given that you are so close to the subject of email as a technology and how people are using it, I wonder if you have any predictions as to how the future of email will play out or any trends that you see developing that would be interesting to call out?

[00:38:46] Christine: There's a couple different things here. So, one is infrastructure side of things. Right now email providers, email platforms, we're just really not designed for the types of things that people are trying to build for today. I see in the future, in a five to 10 year time span, these types of more modern features being built directly into email providers so that you don't need to have these intermediate steps of, "We need to take all this data and translate it into a totally different form so that we can use it for something useful."

Second is, built-in abilities to connect parts of your mailbox to other applications. It's essentially all or nothing right now, you can either give some app like full access to your mailbox or not at all, which is not a great place to be for security. It's because when email was designed, no one thought that you were going to want to connect parts of your mailbox to 20 different applications. Email was for sending communication person to person, but email has changed a lot in the way that people are using it. When email was first designed your mailbox might have had a limit of like 10 megabytes or something like that or less way back in the day. Now, there's so much wealth of information in there that I think that the data is going to be key going forward as to what email turns into.

On the user experience side, we don't really know what that will look like. What we're excited to do as a platform is to enable people to experiment and try many different things, and then build the tools that make it so that many people can build those things easier and faster because everyone needs the same things at the end of the day. Because there's so much interesting data here and it's all very unstructured like machine learning in AI technologies will become important at some point because you essentially need to pull the signal out of the noise. We don't know exactly how that will look and excited about working with folks who want to figure it out.

[00:41:18] Tobias: For anybody who wants to get in touch with you and follow the work that you're up to, I'll have you add your preferred contact information to the show notes. With that, I'll move us into the picks. This week I'm going to choose Trello, which is a tool that you can use for being able to easily track either to-dos or process workflows or you can use it as a project management tool.It's just a very flexible platform and I've been able to get a lot of value out of that. I largely use it for managing my podcast interviews lately. It's been valuable for that. Anybody looking for any task tracking or being able to have sort of just put ideas, it's a great place for that. With that, I'll pass it to you, Christine. Do you have any picks this week?

[00:42:04] Christine: Yes, a little bit random or on a different note. There is this really cool website that just launched today, it's called Founders for Change. The basic gist of it is it's a compilation of different founders at many different companies who are committed to improving diversity inclusion in the tech industry. If you have a company or thinking of starting a company or looking to join a company that cares about diversity inclusion, you should check out the website. 

[00:42:42] Tobias: Well, I want to thank you for taking the time to talk to me today and discuss the work you're doing at Nylas. It's definitely an interesting platform and one that appears to have a lot of useful potential. Thank you for that and I hope you enjoy the rest of your day.

[00:42:56] Christine: For sure. It's great to meet you, Tobias and enjoy the rest of your day as well.

[00:43:00] Tobias: Thank you.

About Author

Tasia Potasinski

Tasia is the Head of Marketing at Nylas. In her free-time, she enjoys drinking soy lattes and reading good books.

Subscribe to Engineering Blog Updates

Start Developing Today

Connect up to 10 accounts (email, calendar, and contacts) for free today.

Start Free Trial