inferential Goto Github PK

I am a Data Scientist, committed to continuous professional development.
In my view, effective data science relies on simplicity, efficiency, streamlined processes, and unified workflows.

Julia and MongoDB. My supporting arguments for this decision:

• Julia is fast, (optionally) typed, reproducible, energy efficient
• It has a wide variety of thematic organizations
• It comes with a very powerful REPL
• Multiple dispatch offers great flexibility and high performance
• Units of Measure are easy to use
• It has an active and growing community
• AOT compilation will soon be possible
• Scientific ML complemented by SimpleChains.jl for solving NeuralODEs on the CPU

Finally, MongoDB is easy to install, setup and use locally or on the cloud.

Had I not found Julia to offer such a pleasant and efficient development experience, I would have chosen Rust:

• A compiled, performant, reliable, immutable
• It comes with great tooling incl. a REPL
• It offers access to great ML and DL libraries
• It has an active community
• It comes with Units of measure
• Learning resources are excellent
• Energy efficiency is also a major plus
• The Scientific computing community is embracing it

fast.ai Practical

• ML
• Linear Algebra
• DL
• NLP
• Data Ethics

Other sources

• 3Blue1Brown
• ritvikmath
• DeepLearning.AI
• NLP Demystified
• Romeo and Julia, where Romeo is Basic Statistics
• Introduction to Statistical Learning (ISL)
• Statistics Done Wrong
• Computational Thinking | MIT 18.S191 Fall 2020
• Working with DataFrames.jl Workshop part 1 & part 2

Many of the materials above are intentionally foundational