I completed data science studies because I wanted to make my e-learning startup more profitable. I am also familiar with auto ML software from giants such as Google auto ML and IBM Watson. Yet I stacked full codes for Hazlo. I still have painful memories of letting AWS auto ML run without bandwidth constraints and a big bill at the end.

Hazlo bills itself as a solution to SMEs and this positioning is spot on. Now small businesses can harness the power of machine learning previously only accessible to big companies and level the playing field. Small datasets can be used and within minutes you get good business intelligence data. Now everyone can be a data scientist for a day. Orja and his team know their stuff and I must admit to bombarding them with all kinds of questions that only a data scientist would ask. I find myself nodding in agreement at their answers and I moved from one code to full-stack.

Now for real data scientists who prefer traditional Python or R coding, I feel you. Rather than go through rounds of tedious EDA and building models, only to be disappointed with metrics, and return to more rounds of iteration, perhaps, run Hazlo first to get ballpark figures, then build models afterward if the chips look all right. Or maybe let Hazlo run the show. I compared Hazlo to results from the usual iris, diabetes datasets, and also the latest Kaggle stroke prediction one to other auto ML software. Hazlo matches them in terms of results. You also get to deploy models in the form of a website where you can key in parameters and Hazlo spits out a prediction. So now we have a complete pipeline from EDA to deployment in minutes. Much better than leaving code in computers that never get deployed at all!