Nordic Semiconductor Thingy:53 brings dual-Arm Cortex-M33 processor power and Edge Impulse experience out-of-the-box 😍 nordicsemi.com/News/2022/06/N…

Finally, this Nordic Semiconductor nice-looking dev board sees the daylight! And also comes with Edge Impulse "preinstalled" :) nordicsemi.com/News/2022/06/N…

Tools are sometimes bottle neck of the (development) process. If you do not have the right tools build them around existing tech and they will speed you up 🏎️ cc Institute IRNAS

Especially true for this moment in time... 🔩

Running this stuff on embedded seems like a sci-fi movie! 👾

Thingy:53 is a great example of a development device that can be used to validate PoC. When I worked on this I really liked how Nordic Semiconductor did a case for it! Also Edge Impulse works out of the box for a superb user experience 🦾

after a year here is a new blog post. #coredump >>> blog.vojislav.me/2023/02/23/cor…

Been to one of those last year, and I can definitely recommend it!

I've been super excited to see the momentum around on-device AI model porting from Georgi Gerganov, Nouamane Tazi, Kevin Kwok, and tons of others. Here are my top tips for newcomers after four years of building in this space! situnayake.com/2023/03/21/nn-…

Almost all founders learn brutal lessons during the first year, but some learn them much more quickly. Obviously those founders are more likely to succeed. So it could be a useful heuristic to ask, say 6 to 12 months in, "Have we learned our brutal lesson yet?"

Were you unable to attend one of our Edge ML Series stops in person? Not to worry. We're bringing the experience to you — virtually on July 19th! Register now: hubs.ly/Q01ThPvr0

What an exciting year it’s been for edge AI, and we’re thrilled to come together at #Imagine2023 to talk about all of it! Join us on September 27th for a free, one-day virtual event to explore the latest innovations in this fast-growing field. Register: hubs.ly/Q01X6QMM0

I would love to hear more about the technical decision to put micro USB 3.0 on the board designed in 2023...

now you can use NVIDIA TAO models. Very cool!

"At the tiny edge, when you build products you are constrained by cost; you cannot go wild,” he said “There is a nominal price point, and it’s between 1 and 3 bucks… and part of the cost is captured by the non-volatile memory in the microcontroller.” rb.gy/gt9kt

Memory is expensive.

I'm so happy to announce that #NLnet decided to fund the development of an open hardware power profiler + battery simulation device I proposed, called JellyfishOPP. More info (and future updates) about the device available here: hyperglitch.com/articles/jelly…

With our new NVIDIA Omniverse extension, you can now use synthetic data generated in Omniverse Replicator to train your object detection model. developer.nvidia.com/blog/how-to-tr…

This combination of nRF5340 (dual Cortex-M33) and nRF7002 is quite powerful and there is a plug-and-play experience with Edge Impulse. Now you can select Thingy:53 + nRF7002eb option directly in Edge Impulse deployment and collect data remotely! docs.edgeimpulse.com/docs/developme…

What do you do for a hobby: I dig through Zephyr Project source files.