3 PHASES OF A DATA SCIENCE PROJECT
All the data science projects implemented using Big-AI consist of three phases (the score here is proportional to the maturity of the project and also indicates the number of customers involved in development)
The 0-1 phase is the first phase
In this phase, we built a predictive model in an ad hoc environment that can predict a quantity of interest for one of our customers. We interacted with our customer-facing teams, product management, and back-end engineers and acquired the right data assets for modeling, performed data QA, understood the problem
space, and ensured that we were experimenting in the right direction. At this stage, we also did a lot of brainstorming, literature review, and algorithm prototyping to explore the landscape of the solution space.
The 1-5 phase is the second phase
This phase involved taking a working ML prototype and generalizing it across multiple customers. To do this, we thought infrastructurally, then trained and evaluated our model efficiently and accurately. We considered what types of feature engineering and algorithmic techniques would allow us to scale.
The 5-100 phase is our third and last phase
This was when we took a promising ML pipeline and prepared it for global release in our software. In this we had done extensive unit and integration testing, prediction stability monitoring, and documentation. We also evaluated and modified the development to optimize cost of goods sold (COGS) and operational reliability. During that period, we interfaced with teams outside of engineering, including product marketing and sales, to educate the rest of the organization on our new capabilities.
For further reference, visit our channel and refer the video named 3 Phases of a Data Science Project or click the link.
For any questions?
Share This :