Katana, modern manufacturing & inventory software provider, has just recently raised $11M in funding. This together with the e-commerce boom will significantly fuel the Estonian SaaS startup business growth in 2021.
But before that, they are actively searching for new katanauts (that’s what they call themselves) to conquer the world with. 🚀🚀🚀
This left us wondering how the everyday work and life at Katana looks like these days?
We talked with Priit Kaasik, the co-founder and CTO at Katana, and asked him tonnes of questions. The conversation went deep into their technological backbone (full stack or function based delivery? in-house development or external services?). Also, the open discussion revealed a product mistake and how heavily underused their office PS5 is.
Take a look at the interview below 👇 and get a glimpse of Katana’s mindset. Or as Priit puts it “Why become a unicorn when you could be a unicorn 140 times over?”
You have jokingly said that “Being a blank canvas and a bit naive” is actually useful when leading the engineering team. How would you describe yourself as a CTO?
There is no good way to hide that I am a process guy first and technology guy second. I aim to be the facilitator of technological advancement to best serve and innovate Katana’s manufacturing platform as a service.
Being a hybrid describes me well – excelling at combining various aspects, never in a singular discipline like coding. My strength lies in defining and driving towards the technological impact that needs to be achieved (I hope others would confirm this).
That said, my favourite pass-time is assembly and tinkering with small-factor but powerful gaming PC-s. I assemble about 2 per month for friends (for their kids) and for people who know this hobby of mine. Some may consider me a gamer, but I‘m really hanging in there because of the technology…
How would you characterize the fundamental values of Katana’s development team?
We emphasize the importance of putting on the end user’s hat and understand the business better – using full stack development and User Stories. Other main pillars would include:
💡 product delivery in increments and iteration,
💡 never reinventing a wheel,
💡 the meaning of “minimal viable”,
💡 writing cloud-native software that is robust and scales with the business.
Mainly, the development team is on a quest for a smart solution.
Let’s talk about Katana’s technological backbone. What stack do you use?
We are currently focusing on adopting Kafka more widely (lessening the importance of internal API to serve dependencies) to scale the service better horizontally.
Micro frontends are something we concentrate on – to improve parallel delivery and deployment from different product teams. And as a product, Katana has set our eyes to solutions that help us to become a manufacturing platform with a popular app store.
Katana uses Heroku, which limits people who enjoy deployment pipeline building. What kind of developers would be a perfect match with Katana?
The limitation is similar to any other choice of technology. It’s just a combination of Bitbucket Pipelines and Heroku Pipelines – a Katana flavor of it. It would greatly benefit if you:
✅ Like pattern-based engineering, building Lego blocks for yourself, your team and teams.
✅ Appreciate full stack development and the need to understand a customer.
✅ Enjoy working in an Agile team.
✅ Being a self-sufficient engineer, without the need to be actively managed and supported, is a treat we value a lot. Because we really prefer to manage people as little as possible.
A person, who enjoys tinkering and overpolishing solutions, will have a hard time at Katana. We look for engineers who understand that their value lies in their ability to dish out the “right stuff at the right time”. Not in the code worthy of a monument now and 10 years later (we value good code but not overdoing it).
What aspects are not developed in-house, but bought as an external service?
Let’s say that the scope of what Katana is doing, is roughly described here. And in order to deliver it, we use following services in Katana:
- Auth0 CIAM (soon to be renamed to Okta for User Management)
- Recurly (and Braintree) to manage subscriptions and payments
- PDF Generator API to create PDFs, of course, invoices, bills etc.
- Pusher – In-browser notifications and data updates
- Ag-Grid – React Grid
- Intercom – In-product Support Chat
- Elev.io – In-product tooltips and knowledge base
- Fullstory & Hotjar – product analytics
- Appcues – for onboarding
It is also important to note that Katana doesn’t have any technical operations nor DevOps people on-board. We consume data (Postgres, Redis, Elasticsearch, Kafka, Bull etc.) also as a service.
The cat or dog person question. Full stack, or front end & back end separately?
Full stack. Regarding customer features and end-to-end transactions, it’s far superior compared to function based delivery.
I have observed and participated in such designs in various top notch software development companies. In the case of web-based SaaS, I would always choose full stack to win it at the market.
It’s a focus thingie – eventually a product team has to be full stack and cross-functional, you have to work with the engineers you are able to hire.
In which case would you consider hiring front end and back end specialists instead of full stack?
To compensate for a weakness, bolster skills that are lacking or not represented enough in the company, but are important.
Looking back, what were the technical mistakes that you wish that could have been avoided when building the product? How did you fix it?
Just one thing pops to mind right away. We decided to upgrade our backend node.js framework LoopBack3 onto version 4. It has turned out to be a very expensive work with very little benefits (other than LB3 is no longer supported).
Should have assembled our own modification from a more lightweight backend framework a la NEST.
What direction are you scaling (horizontal/vertical etc)?
In all directions, of course!
But seriously, in the engineering team we emphasize the importance of horizontal scaling and the code supporting it.
Vertical scaling is for the emergencies, when the performance need is miscalculated and a performance issue quickly needs addressing. That is another reason why we chose Heroku, it really excels at this type of performance management and scaling.
Let’s imagine, I’m a newly hired engineer. How will my first days at Katana look like?
Once we agree and assemble the work gear and software set with the engineer – wintel, mac and/or Linux – you will be ready to go. You choose between onboarding in the office or remotely (your gear will be delivered at home). We use Slack and Whereby to communicate and run meetings etc. Also, this setup works really well with onboarding.
We also have a compensation system to expense the cost of home office setup and including headphones with active noise cancelling (must-have product these days).
When you’re comms-capable, the first weeks are for meeting fellow katanauts – learning the heartbeat and key information sources. In software engineering our current gold standard benchmark for the first day is to get the development environment up and running by noon, join our food club, have team lunch, fix a bug and get the fix to live by the end of the first day.
Fully remote or office?
Nobody is working fully in the office nowadays. That said, we are eagerly waiting for situation improvements, so we could restore the office experience to its former glory.
Even during the pandemic (when approached with precaution), our office has its perks (heavily underused PS5, barista-grade coffee etc). And our lunch club is still functioning! About 15% of people use the office at the moment, wear masks, and have spread out. Our office is new, roomy and really well ventilated.
Actually our new office will be ready by summer. It will be located in the newly renovated building next door to Põhjala Taproom at Noblessner. Even with new ways of working, we decided that having a joint workplace and meetings in person will remain an important part of our culture.
Katana recently raised $11M in Series A. What are the main plans for the engineering team in 2021?
Double up the size of the engineering team each year!
The aim is to build a team to become a world’s leading manufacturing platform able to challenge SAP and such. You cannot really catch them otherwise – why become a unicorn when you could be a unicorn 140 times over?
But (there is always a but), business also has to grow accordingly, and so far we have mostly been hampered by lack of features to fully tackle SMB/SME manufacturing space.
Sounds like your cup of tea? Check out Katana’s current openings and apply!