Workshop time! 🤓 Spending the afternoon at a fantastic virtual session learning about Clinical NLP.
More info here: https://t.co/ylPaMkGR2c
Zoom: https://t.co/p5xCVVyWps
#ClinicalNLP
Module 6 of #MLOpsZoomcamp done!
I tackled best practices for reliable & maintainable ML systems:
• Refactoring code & writing unit + integration tests with pytest
• Enforcing formatting & pre‑commit hooks
• Automating CI/CD pipelines using GitHub Actions
#MLOps#pytest
When your ML model drifts silently, you’re not deploying — you’re gambling. 🎲
Just wrapped up the #MLOpsZoomcamp monitoring module:
Logged prediction drift & quantiles with Evidently
Stored results in PostgreSQL via Python
Visualized everything in Grafana
#MLOps#DataScience
🚀 Just wrapped up another #MLOpsZoomcamp module!
Learned how to deploy ML models using different strategies:
🔹 Online (Flask API, streaming with AWS Kinesis + Lambda)
🔹 Offline (batch processing)
Now models aren’t just trained—they’re serving! 🍽️
#MLOps#MachineLearning
Just started orchestrating my ML pipelines with Prefect — and I’m loving the clarity it brings!
Here’s a peek at my flow run from this week’s #MLOpsZoomcamp 🔍
Automation + observability = 💪
#MLOps#Prefect#MachineLearning#MLFlow
Week 3 of #MLOpsZoomcamp complete!
Dove into orchestration with Prefect and learned how to track & manage models using MLFlow.
Now my ML pipelines are not only automated, but also reproducible and trackable!
Loving the hands-on approach 🙌
#MLOps#MachineLearning#Prefect#MLFlow
The 2022 WFBMI Colloquium has started! Dr. Metin Gurcan, PhD is giving a quick introduction and highlighting the accomplishments of the Center for Biomedical Informatics. @WFCTSI@AtriumHealthWFB
More than a decade ago Dr @metingurcan and I proposed the idea of a #DigitalPathology conference at @SPIEevents Medical Imaging. In a perspective piece in J of Med Imag, we discuss conference history & impact on #AIinPathology#computationalpathology.
https://t.co/1ze5g0grud
Exploring the Sharpened Cosine Similarity Layer I was able to reach 99% in MNIST with only 3,140 parameters. I think that is a record. Shows the potential of the Layer in low parameter settings. Not sure yet if and how it scales up though. Still exciting!
https://t.co/AiFJAZKTgY
Today we say goodbye to our #WFBMI 2021 Summer Interns. Lots of great project work presented today, their futures look bright! Congratulations to Ben Morris on his session award!