Getting into AI research has been one of the best decisions I've made.
We're now able to deploy @spairally 's memory-efficient models across servers, phones, and microcontrollers ; with seamless integration and offline inference support.
Build once. Run everywhere.
The debugging I have debugged for spairally's models to run on a microntroller ??
And its just dependency issues -
I had forgotten how much I dont like python ...
Today's my last day this month at GTC ...
1. Offline features - DONE
2. Business meeting with an angel investor - DONE
3. Taking a mirror selfie - DONE
4. Sitting around enjoying the office - in progress
Not in the mood for hardware research today - really !
The city is ever beautiful from here …
Getting that office at GTC … one of the best decisions I’ve made in the recent past .
And I’ll tell y’all why soon .
Reducing false positives was our priority with @spairally app - from the word go .
As a result :
1. False positives ( unheard of since launch - except for the 15 minutes window that we were swapping the api )
2. Extremely strict validation during pre-processing …
Another time I was puzzlePrincessOfNairobi👸 …
And there was almost no puzzle I couldn’t crack . Of 5000+ puzzles , I couldn’t solve only 2 .
This was one of the two .