DEV Community


Posted on

A Road From Software Engineer to Machine Learning Professional via the Core Designer Learning Path of DataIKU

In my lengthy career on the web, 25 years, I have seen and been part of the evolution of web 1, 2, and the emergence of web 3. Prior to joining Fourteen33, I had many technical lead titles that required that I worked as a "Full Stack" engineer and work with "lots of data".

During my journey of learning not just the basics, but a practical work flow of ML projects, I found an accelerated learning path of technical terminology and entry points thanks to using Data Science Studio built by DataIKU.

The further I worked through the certification and deeper learning modules of DSS (Data Science Studio) the more I realized just how robust DSS (Data Science Studio) was in terms of integration with our numerous back ends and servers that we traditionally use for large scale Machine Learning Projects.

For instance, Spark Integration, Custom SQL Connectors, PostgreSQL integration, and caching csvs from remote data-sources were all very interesting data-points in helping me better understand how DSS operates as a full platform capable of "interop" with many other server and data technologies.

Another interesting feature I liked a lot in DSS is the ability to quickly build "recipe" transformations into new datasets (views) of the original datasets without affecting the original datasets source. With lots of "web development" experience in my background, I was pretty impressed to see the in built functions libraries like "Extract Geo" from a column with IP Addresses, or "Extract Header Data" from the long header information string stored in a column, neatly parsed into clean columns for analyses directly after running the inbuilt functions.

Whether you use the entire DataIKU DSS platform, or just a portion of the data preparation tools, its clear there are projects that can benefit from using DSS along side many of your other ML technology systems, including GCP, Azure and AWS.

Over-all, a key experience I gained from the Core Learning Path Certificate experience, was learning the common language so that I can talk at an intermediate level with our seasoned Data Scientists and Data Engineers.

For a quick way to kick-start your learning path with DataIKU Data Science Studio (DSS).

An option I chose was to open an Google Cloud (1 year 300 dollar) credit account, and spin-up a DataIKU server. Note: before deploying your DataIKU server, you can actually reduce the benchmark Virtual Machine requirements to its lowest allowed settings to help save money on Compute Time. Even though I ran my test server on a "free GCP" account, and it estimated to use about 50 dollars per month in usage credit, I still would start and stop my machine when not using it to save credit. One other point, you can install DataIKU server on your local machine (any flavor of OS is supported).

Here is the link to my Core Designer certificate which is good for 2 years.
(in all I probably invested about 15 hours of learning time including the final tests to achieve the certificate).

I have such a better grasp on the full lifecycle of some core types of machine learning projects thanks to DataIKU, I am already enrolled in the next certificate course named "ML Practitioner".

To close out this post and journey, and on to new adventures, it should be stated again.. I have years developing in many language types for the web. Webmaster, Architect, Front-end dev, Teacher, Full Stack... all of those titles and more have dotted my resume in my career. I think we all now know the superb need to have strengths in Machine Learning in AI and how its applied to the modern commerce and business verticals. Give DataIKU a try, and you will accelerate that learning path a great deal.

  • Happy Coding!

Top comments (0)