Recent Posts

• March 08, 2022

  Takeaways from the NeurlPS 2021 Conference

  My experience with NeurlPS 2021 Conference was phenomenal despite I have to join the Zoom meetings from a shitty quarantine hotel at Shanghai as if I began a prison sentence.Anyway, following is the main takeaways from some of sessions that I part...

• October 30, 2021

  What I did to improve the data pipeline efficiency by 100x in MS SQL

  I was going to build a training data pipeline fetching fresh labelled data from a MS SQL DB continuously. Soon after I noticed scalability would be a bottleneck here since this MS SQL DB is very OLTP focused and not indexed for massive analytical ...

• August 02, 2021

  Get Started with Dask

  Dask is simply awesome. If you are looking for Spark alternatives that are light-weight, simple, easy to set-up, native to Python, and do its job (i.e distributed computing) well without changing much of your existing Python codes (in numpy or pan...

• September 11, 2020

  Constrained Contextual Bandits for Personalized Recommendation

  Contextual bandits (CB), unlike the full Reinforcement Learning problem in which there are environmental states, new states depend on previous actions, and rewards can be delayed over time, “contexts” under the CB context do not mean the environ...

• August 09, 2020

  Democratizing Building ML Web App with Streamlit and Heroku

  R ecently I integrated a TensorFlow-enabled ML model into the Streamlit web app in lessthan 40 lines of codes. Then I deploy the app on Heroku,a low-code platform for deploying, managing and scaling modern apps. In this post, I will show briefly h...