first failure of 2026: solving 100+ deep learning questions
My only resolution for 2026 was to solve one machine learning problem a day.
Made it 20+ days. Went to bombay for concert and lost my streak… I did see Linkin Park tho :P
Still solved 100+ problems though.
why i’m done
DeepML is solid, but:
- Test cases suck
- Runtime is painfully slow
- The “1 problem/day” pressure was killing the fun
new resolution
Chill the f out:
- Papers -> Implementation
- TensorTonic for practice (shameless plug)
- CUDA Reading CUDA for Deep Learning one chapter / week - because GPUs rock
All my from-scratch implementations live here: ML Repo