Many people learning AI/ML feel stuck because of ๐ฒ๐๐๐๐๐๐ ๐๐ ๐๐๐๐๐๐๐๐๐
consuming so much info that they donโt know where to start, what to focus on, or how to apply what youโve learned.
First step is to know your career path.
Lock your doors, switch off your phone and cry out to God in prayer.
Pray.
Push in prayers until you feel a release.
Then watch what happens in the following days.
I thought it was a popular opinion!
One more advice:
-Don't die in Jupyter Notebook
Move on to Deployment, scaling, MLOps, and MaaS integration.
get comfortable with cloud & infra.
End2End skills will make you far more valuable than someone who only knows model building.
math & statistics
โ
python
โ
classical ml
โ
advanced ml
โ
nlp
โ
deep learning
โ
transformers
โ
langchain
โ
langgraph
โ
mcps
โ
finetuning
โ
fastapi
congrats you are an ai engineer now
I saw this two days ago, and I've been inspired. I asked my advisor why we don't use this GAF method on our pediatric sepsis time-series data, and she was skeptical, saying she tried it during her PhD.
Since I'm a delusional optimist, I'm committing 40 hours to experiment with a GAF-CNN-BiLSTM architecture. beast.
Best case: It closes my AUPRC gap to 0.83, and I submit the paper to ML4H 2026.
Worst case: I publish the exact ablation study explaining why this specific architectural risk failed. The science wins either way.
I thought it was a popular opinion!
One more advice:
-Don't die in Jupyter Notebook
Move on to Deployment, scaling, MLOps, and MaaS integration.
get comfortable with cloud & infra.
End2End skills will make you far more valuable than someone who only knows model building.
Unpopular opinion.
ML engineering is more software development than Data science.
In conclusion, itโs better to get into ML engineering from dev than DS ๐
Iโve heard people complain about the algorithm, now itโs my turn๐คฆโโ๏ธ
Lately, My feed been flooded with too many random tweetsโฆ i didnโt show interest in any form.
@nikitabier@elonmusk this new algo needs a retouch.