Varadharajan Raman is the Principal - Value Engineering at Flutura Decision Sciences & Analytics. He enables digital transformation pursuits of manufacturing enterprises by leveraging new age Digital technologies and fast forwarding the Industry 4.0 realization. Also as Head of Energy Practice, he works with key stakeholders across the energy value chain to leverage digital technologies and improve overall operational efficiency.
With industrial systems getting increasingly sensorized, high frequency data has become the modern language of machines in this fast evolving digital world. This has led to a paradigm shift in the world of “Domain Natives” (maintenance/ reliability / process / operations / quality engineers) with data diminishing the distance between them and their assets or processes, even in case of remote operations.
Hence making sense of large volumes of high frequency sensor data has become a pre-requisite for the “Domain Natives” to manage operations effectively and make right decisions on time consistently. For example, when a machine breaks down, the maintenance engineer is expected to discover the root cause of the failure analyzing the time-series data from the sensors before dismantling the damaged components of the machine. This creates (positive) pressure on the maintenance engineer to double-up as a quasi-data scientist to handle the new normal.
As a passionate tech professional operating at the intersection of Data and Domain, I get a panoramic perspective of personas coming together to apply data science for a given domain problem and create business value. Being an obsessive math enthusiast, I visualize the combination of technical capabilities required as a continuous spectrum with vertical specific domain proficiency of “Domain Natives” on one extreme and horizontal data science competence of the “Data Natives” on the other extreme.
Given the orthogonal relationship between the Vertical specific “Domain Natives” and horizontally driven “Data Natives”, the true north of Industrial AI is about connecting the two extremes of this spectrum at scale.
One of the major challenges in establishing this connect is the linguistic barrier between their technical vocabularies. With deep rooted domain specific problems, it becomes a herculean task for an external “Data Native” to contextualize and apply data science. This leads us to the critical question - “How to make the “Domain Native” self-reliant to perform their analysis independently?
The nature of work of Domain Natives is full of technical decision making across the range starting from simpler decisions like changing set-points to complex ones like identifying the root cause of an equipment failure. But most of these decisions are driven by a combination of experience and subject matter expertise. While experience and operations knowledge are great technical assets of the Domain Natives, the ability to analyze high frequency sensor data independently can be a great catalyst to improve their quality and velocity of decision making.
In my opinion, the real measure of success of Digital transformation is about how much of these day to day decisions are made accurately, accelerated and automated through precisely calibrated decision support systems leveraging data at scale instead of relying on the tacit knowledge of skilled domain natives.