Does anyone here work on implementing Machine Learning into their software application for their job? I have a few questions about the field.

I have considerable experience in the area. Feel free to PM me.

I would say that the work of Machine Learning engineer can be divided into three major chunks :
1. Almost all Machine Learning algorithms have implementations available, one important task is to pick what algorithm is to be used for a task. (for example decide whether DBSCAN or K-Means is more suitable for a clustering problem, or which out of MaxEnt/CRF is better for NLP tagging).
2. Feature Engineering is another important part of the job. The raw data generally cant be fed to the algorithms above directly. This might require performing analysis over data, pre-processing it, dimensionality reduction and things like that. This takes the major chunk of time.

The combination of two tasks, reiterated many times, finally leads to good results. We might need to pivot on algorithm and feature extraction many times .


3. Staying abreast of latest development is also important. This field is one of the hottest research areas and newer and better algorithms are being invented. A major amount of time goes into reading too.