June 05, 2023
 |
For me, GDSC was a kind of starting point for this adventure. As an introvert, I initially struggled to connect with others, but being part of a larger group motivated me to step out of my comfort zone. I gained confidence in myself and my ability to make a positive impact on others.
What I love most about Google Communities is the emphasis on human relationships. Collaborating with other leads and learning from one another is integral to driving our communities forward. Even now, I enjoy the ongoing exchange among alumni.
I'm an Android Developer by profession and I use Google tools on a day-to-day basis, starting from AndroidX libraries, Google Play Services, Jetpack Compose and more. At a company where I worked previously, I spent a lot of time working with Firebase and Google Cloud.
Jetpack Compose is hands down my favorite tool to build an Android app since it provides an intuitive way of building a user interface that saves me time and effort. Before that, I was not really friends with UI designers. Creating UI was a chore, but Jetpack Compose is a game-changer.
I’ve been working on an application called Be Served at Zx Connect. The goal is to make life easier for users by offering services like goods and food delivery.
The app is entirely written in Kotlin and uses modern tools like Jetpack Compose, architecture components, and Google Cloud Platform, which help us build scalable solutions. They easily improve and add features based on customers’ feedback.
The first step is often the hardest. The best way to persevere is to find people who are like you, who have been where you are and who can inspire you. The best place to be is in a developer community by attending meetups and being active.
Another important piece of advice is to find a specific area and focus on it. Practice is the best way to understand a concept; learn by doing, and never neglect basics.
Becoming a better developer takes time and effort: stay passionate and trust the process.
We have a responsibility to educate ourselves and others about the issues facing marginalized communities, to speak out against discrimination and prejudice, and to actively work towards creating a more diverse, inclusive, and equitable workplace.
The role of allyship in the tech industry during Pride Month is crucial, but it should not be limited to just one month. It's about creating a culture of acceptance, respect, and support all year round.
 |
I'm currently a big fan of the Workbox library. I've used it to build all my recent progressive web apps, which require core PWA capabilities like offline readiness. I've given a lot of tech talks and workshops on Workbox because I want to introduce it to as many developers as possible.
I strongly believe that installability, connection resilience, and proper integration with the underlying platform are essential components of modern web frontend applications. Building a service worker from scratch to provide these features can be challenging, as there are many nuances and things to consider to avoid ruining your app in production.
Workbox provides the perfect balance of abstraction for the core entities you need to optimize networking, while still giving you full control over your service worker behavior. It's a stable and mature library that's actively supported by the community.
In my 20 years of being a developer, I've used many tools from Google. But what really drew me to the communities, conference organizing and technical speaking was the Angular framework. I was instantly in love with the technical side of it - it was a huge step forward for web development, allowing for the creation of frontend apps using some really cool backend techniques.
I also appreciate the developer community around it. They focus on sharing knowledge and tools, helping out new people, and being welcoming and positive. After attending a few Angular events organized by the community, I became a co-organizer of the Norwegian Angular meetup, and co-founded the Nordic Angular conference: ngVikings. We had three in-person events in Denmark and Finland, and one online event with thousands of participants. Nowadays, I use multiple frameworks for my hobby and side projects, but I'm still an active member of the Angular ecosystem.
My recent projects that use Workbox for the service worker automation include:
I built it for my own live performances and share it with the global community of DJs and music producers.
- A Web Push API playground for the developers.
Web Push is one of my favorite features of progressive web apps. I actively explore it from the development and UX perspective, and present my findings.
- A Conference about Prompt Engineering as a discipline.
I am an active member of the generative AI community. This is my way of contributing to the creation of a healthy and united community around AI and LLM fields.
They are all open source, so I would appreciate any comments or pull requests!
 |
My favorite Google tool is TensorFlow, as it has extensive support for a wide range of applications, from tabular modeling and graph neural networks to computer vision or natural language processing. The ability to build, train, and fine-tune complex neural networks using TensorFlow has significantly accelerated my research. Its easy integration with other technologies has made it an essential part of my development process.
Also, I love serverless and use Google Cloud Functions in nearly all my projects! I'm excited that during I/O this year, Python support was introduced in Firebase Functions as well.
I have used a variety of Google tools in my projects, including TensorFlow, Google Cloud Platform (GCP), Firebase, and Flutter. They enable me to create powerful machine learning models, deploy, manage them at scale, and to build user-friendly interfaces to serve their applications across all platforms.
One notable project I have built using Google tools is the Bulgarian NLP platform "AzBuki.ML". I used TensorFlow to develop several state-of-the-art natural language processing models specifically tailored for the Bulgarian language. These models were hosted on the Google Cloud Platform and served through web and mobile applications built with Angular, React, or Flutter and hosted on Firebase Hosting.
In the past two years, I have been actively conducting research in machine learning as well. I have used the developer resources by Google to extend my theoretical studies to usable software libraries or at least provide tutorials to interested developers on applying state-of-the-art techniques for auxiliary learning, contrastive learning, tabular modeling, and autoregressive text generation in their work. These include:
- The GatedTabTransformer: A state-of-the-art deep learning tabular classification architecture inspired by TabTransformer with integrated gated multilayer perceptron. I recently gave a talk on how it can be used together with the TF-DF library for robust tabular classification and regression.
- RSTOD: Novel auxiliary tasks for task-oriented dialogue systems. The study has been peer-reviewed and is available in the ACL Anthology.
- Start your developer journey with curiosity and continuous learning. Resources for computer programming and machine learning are easily accessible, allowing growth at your own pace, from anywhere, at any time, and at any age.
- Use online courses, tutorials, forums, and blogs to learn and connect with developer communities.
- Collaborate on projects, exchange ideas, and gain insights, support, and mentorship from these networks.
- Work on open-source projects that you're passionate about to enhance your problem-solving skills through hands-on experience and engage with the developer community to gain valuable experience or forge connections.
By embracing allyship, we can help create a more inclusive and innovative tech industry that benefits everyone.
Posted by