2022 Retrospective

It’s almost mid-February, so I am pretty late with this, but better late than never I guess. I do enjoy writing my yearly retrospectives (you can see my previous ones here) since they give me an opportunity to pause for a second and reflect on the work done, something I usually don’t do as I am always running to the next thing I want to work on.

Web3 and NFTs

My retrospectives for the last few years have begun with the myriad of technologies I have learned that year, from VR/AR, to chatbots and Machine Learning. This year was a bit different though. I finally got to pursue a technology that I have been passionate about for a long time (and have written about a number of times previously) but have had a hard time finding a project relevant to my company: Blockchain and Web3.

In 2022 I spent most of my time developing and pitching a Web3 idea that resonated well internally. I am really excited that the effort is ramping up in 2023. The Web3 space is a curious one and is often a target for ridicule but the potential for creating experiences where end users are not just passive consumers but active participants in the evolution of an experience is pretty exciting. I wrote some of my thoughts on why NFTs are an exciting space and worthy of a second look, but even if it doesn’t immediately convert you into a believer, at least be aware that a lot of very smart people are working with more noble goals than get-rich-quick schemes.

Building Teams

My role has often been at the early stage of products that then ramps up to a full production team. This year I did quite a bit of that as well, ramping up the UI team for a soon-to-be-announced product. The role involved helping migrate an internal prototype to production quality code as well as hiring a team that would take it forward.

While I have interviewed candidates for small teams before, this was for a larger team with folks being hired for roles at various levels. I am pretty proud of the team we finally built but I do want to grow my own internal framework for assessing engineers and managers.

I have seen 2 approaches to hiring:

  1. Have a loose idea of the skills you are hiring for and then find the candidates that are close to those. If the matches aren’t close, still hire for the skills you find.
  2. Have a very exact idea of what you are looking for and seek out those who match that very strictly

    Most approaches fall somewhere in the middle of those 2 extremes but sometimes we lean too heavily on #1, especially when we are in a rush to hire candidates. One approach I found useful was to explicitly list the technical and cultural skills I was looking for in an Airtable table, develop a set of questions that would reveal those ahead of time, and use the table to explicitly score the candidates. It worked well for the most part but I have a lot more to learn myself in that role.

Lightning JS

For both the unannounced product and our explorations of Web3, I spent a lot of time working on TV UIs using Comcast’s open-source Lightning JS UI framework. LUI is a JavaScript-to-Canvas/WebGL framework that skips the HTML DOM completely. The approach makes it a lot more performant on low-memory devices like TVs and Set-top boxes. While the approach is great, some of the patterns in LUI are very different compared to other UI frameworks I have worked with. I am pretty competent with LUI now but it was an adventure to get there.

An interesting approach at work is that higher-order UI components are now being built by a team reporting to the design/UX leadership. This has made the components very faithful to the visual design patterns that the UX team imagines, though I do think there are optimizations at the performance level that need to happen since we do hit weird jank issues on some of the devices we work with.

LUI is focused on TV-based experiences but might be fun to try some mobile/desktop-optimized apps/games that don’t need the native UI frameworks.


In my off-time, I still enjoy working in Flutter. This year’s fun side project was a grocery-tracking app I named Grocery Timer. GT has not revolutionized the world of grocery management but has still been a fun playground to try different UIs for a surprisingly complicated problem space. Keeping things simple so that data entry isn’t a chore while still providing value and the right reminders has been a challenge. That said, I personally find the app very useful and so its relative obscurity on the app stores doesn’t hurt too much πŸ™‚


Didn’t do as much reading as usual this year but did find some very good reads, specially Loonshots and The Courage to be Disliked. Highly recommended. I am also rediscovering the joy of graphic novels and YouTube channels, especially For the love of Comics, has been a great source of great recommndations.

I didn’t write as much in 2022, at least not publically, but that is one of my resolutions for 2023. Stay tuned πŸ“»

Author: Arpit Mathur

Arpit Mathur is a Principal Engineer at Comcast Labs where he is currently working on a variety of topics including Machine Learning, Affective Computing, and Blockchain applications. Arpit has also worked extensively on Android and iOS applications, Virtual Reality apps as well as with web technologies like JavaScript, HTML and Ruby on Rails. He also spent a couple of years in the User Experience team as a Creative Technologist.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: