Skip to main content

TechSpeakers Ahoy! : A Berlin Meetup

Mozilla Foundations is a curious organization. Though most of us have grown up getting acquainted with it by using Netscape and later Firefox, Thunderbird. There is also the foundation part where it undertakes a lot of projects to make the openweb and world as a whole a better place.
Among many teams that help volunteers, budding developers dip their toes into code, help in evangelism, code contribution, advocacy is Developer Relations (DevRel) and they came up with an ingenious plan to come up with a team of volunteers passionate about these to come forward represent Mozilla DevRel. Havi Hoffman (from Mozilla Labs), Dietrich Ayala (From Platfrom Engineering, Firefox OS) and many others handpicked a group of people who eventually came to know as Mozilla TechSpeakers. (If you want to read more about it, this lovely piece by Havi is a must read!). I was fortunate enough to be in the first group (knows as pilot pilots). And eventually ended up being facilitator to the last group which graduated the program just this year!
Mozilla Tech Speakers
Just this August I attended a global meetup of Mozilla TechSpeakers with the participation from around the world of 26 people. It was overall a very tight schedule, and presented us with a unique opportunity to get our speaking hoodie up and present to a friendly and safe audience.
Mozilla Tech Speaker Team, in *serious* mode

The idea was to have all of us present on a topic of our choice on those two days among our peers and get constructive criticism, suggestions and opinion from them. I felt, the fact that you can actually be free form any worries of goof-up and present to actually your friends was a huge bonus for all of us speakers, battle tested and new learners (after all we are all “Moz Tech Ducklings”). Denise Graveline was our coach who assigned task to us. The task for this particular meetup was a 5 minutes talk without using any slide or presentation. Though she had us all informed of the rules beforehand and mostly everyone had been practicing their talks, I, being the in-house procrastinator came up with a wild idea just the night before my talk (which again was the 3rd talk in the morning *sigh*).

And just as you can imagine with ideas like that, you lose your sleep, hack the night, prepare your….demo (cannot use slide) and device a mischivious plan to get around no slide rule. Next day, I arrived at betahaus (an awesome co-working place at Berlin which I would love to visit again!) armed with two mobile phones and my trusty Nexus 9 fully charged. The thing is, we are not allowed slides/presentations and even charts papers, but all I needed was a place to show QR code and a camera (which my mobile conveniently provides!). So I took the dare and pulled off a live demo of a Augmented Reality application demo and an exploit I just cobbled together the previous night to show a live exploit in mobile (and the demo goddess was pleased on me so went without a hitch!).
As Flaki said
Everything went perfectly well, except I had planned people to actually open their mobiles and try to do the demo with me (to see the exploit live in their phones) but people seemed a little bit shy to try an actual exploit (come on people! I was just instantly crashing your chromes and firefox, not like I was doing anything too bad…).
But that messed up my timing and I kind of finished 1 minute 30 (approx?) seconds before (I had also factore QA time, which didn’t happen either). But the demo went without hitch and I was happy! Got some awesome feedback from everyone and a lot of MozLove form Michaela, István, Andrzej, Andre, Michael and everyone else (you all rock!).
You can have a look at the talk (and demo) here!
And did I forget to tell you how Bæla covered us?
Overall this was a learning experience on how squezzeing a talk in 5 minutes seemed so much of a problem than a 45 minutes one I am used to. But when I finally made that to work, I was amazed at how much I could squeeze into 5 minutes (I literally had 33% time free in my talk). That was a revealation.
You finished it in <4 minutes? Really?
Listening to everyone made me realize how most of our battle tested developers and speakers function and their different traits. Not everyday you get to see their own speaking styles up close and get to analyze them. I picked up a lot of *TO DO* from how Andre, Laka, Ioana, Andrzej and Flaki handled the talks. Everyone had their own style, and I realized there is something I could learn from each one of them. After all there is on point if every speaker follows the same pattern, that loses the charm. Also your own content definately helps the most! If its made in your hand, its your story, it’s unique to you! Nothing can bring out your passion and convey it to your audience more than what you have spent your time doing!
And did I show how we normally were throughout the whole day?
Like
Hence...
Except our ever so serious Flaki
But of course since Laka, true to his pose says
How can we leave Christos who mostly needs...
With that we finished our two days shenanigans. And signed off our adventure. Which brings us to our crazy pose!
The True and Crazy Mozilla Tech Speakers!
Want to know more about us?
Why not follow our adventures in our Twitter @mozTechSpeakers and #mozTechSpeakers
We fight and thrive for openweb, we are Mozilla Tech Speakers

And there actually is a little side story to all this, and the talk. So Dietrich swore he will not vouch for me again if I don’t report the exploit immediately, and later in Viewsource when I showed the same exploit/bug to Robert Nyman on Chrome he suggested the same. And finally I have done my part for Firefox atleast!
**All the pictures in the blog post are courtesy Christos. To see more of our pictures and what happend, visit his Flickr album
Labels:

Comments

Post a Comment

Popular posts from this blog

Racecraft (Project Koru) · Prologue — The Origin Story

Racecraft · Prologue , The Origin Story It Started With a Wine List and a Question About Racing How a happy-hour conversation in the Bay Area turned into a trustable AI race coach , and then into a second version that runs entirely on a phone, on the NPU. This is the prologue to a five-part series. Two years ago(1st November, 2024) I was in the Bay Area for a GDE Summit. If you've never been: it's a couple of days of talks among Google Developer Experts, the kind of people who get unreasonably excited about a new on-device runtime, and then , mercifully , a happy hour where everyone stops performing and just eats. We ended up at a restaurant(Puesto Santa Clara), a long table of GDEs, and I was doing the most important engineering of the evening: trying to decide which wine to order. Across the table was Ajeet Mirwani . I don't even remember how the wine talk turned into racing talk , these things drift , but the moment the word "racing" ...
Post a Comment
Read more

The Throughput Trap: Benchmarking vLLM on OpenXLA and the Reality of Production LLM Serving

vLLM Systems · DevLab 2026, Deep Dive I was recently invited by the Google TPU team to speak at the OpenXLA Summer DevLab 2026 . This post breaks down our deep-dive evaluation of the matured vLLM + OpenXLA stack, the fundamental engineering mismatches between CUDA and XLA serving paths, and why traditional capacity metrics are lying to you. If you are operating large language models at enterprise scale right now, your platform architecture team is likely staring at a massive infrastructure crossroads: Should we migrate our core serving workloads from GPUs to TPUs? Historically, NVIDIA's CUDA ecosystem was the only serious option for user-facing, low-latency LLM generation. But here in 2026, the economics and infrastructure options have transformed. Google TPUs are highly available, cheaper per chip, and the open-source serving stack built around vLLM and OpenXLA has officially achieved absolute production readiness. Yet, when our infrast...
Post a Comment
Read more

A Split‑Brain Neuro‑Symbolic Training Method for High‑Velocity Autonomous Coaching from Telemetry

 Author: Rabimba Karanjai Scope: Problem statement + data methodology + model training (no deployment discussion) Abstract Real‑time coaching in motorsport is a safety‑critical learning problem : a system must map noisy, high‑frequency telemetry to short, actionable guidance that remains physically consistent and avoids hazardous recommendations . This paper proposes a “Split‑Brain” training formulation that separates (i) a semantic coaching target (what action/critique should be expressed) from (ii) a reflexive interface (how actions are represented as compact, verifiable tokens). The approach trains a Small Language Model (SLM) in the Gemma family [1] using QLoRA fine‑tuning [2] , and introduces a telemetry tokenizer plus teacher‑student synthesis pipeline to generate instruction‑action pairs at scale. Core contribution: a reproducible method to convert “ golden lap ” differential tel...
Post a Comment
Read more