I’m Dhiman Deb, working as Software program Growth Engineer II at Oracle with the Knowledge and Analytics product growth group for the final 2 years. My complete expertise within the trade is 7+ years, spanned throughout a number of enterprise domains resembling Banking & Retail, Oil and Gasoline, Provide Chain, Telecom, and so on.
At Oracle, I work with the product growth group to develop new options and functionalities to allow clients to have a greater expertise within the Provide chain space, ranging from Order to Money, Procurement, Order administration, put in base, Upkeep, and so on., utilizing numerous instruments resembling OBIEE, Oracle Analytics Cloud and applied sciences resembling Machine Studying, Synthetic Intelligence, Course of Automation utilizing Python, Scala & Spark, and so on. We now have noticed that we are able to higher equip the shopper to inventory their stock primarily based on the expertise with the provider and their supply efficiency.
We now have been utilizing numerous applied sciences to organize an information pipeline and feed the info to analytics functions for Machine studying consumption. Supply knowledge is getting injected from a number of supply methods resembling CSV and Oracle databases into an information warehouse utilizing Oracle Knowledge Integrator (ODI). We now have accomplished numerous knowledge manipulation steps on supply knowledge utilizing ODI inbuilt transformation. Then reworked knowledge obtained pushed into the Knowledge circulate of Oracle Analytics Cloud (OAC), and a pre-built & examined Machine Studying mannequin (Numerical prediction) was used to forecast future demand. Later we visualized the prediction utilizing the OAC knowledge visualization software.
There are two situations significantly: I’ve utilized AI/ML other than numerous small/medium initiatives the place I’ve used python to automate or construct instruments for the group for higher buyer expertise.
Within the first state of affairs, the place I’ve used the BERT mannequin together with my group members to develop an NLP resolution that may present Oracle-specific solutions and hyperlinks to paperwork for numerous analytical and business-related questions.
On the second, the place we try to determine future demand primarily based upon numerous parameters resembling season, previous years’ demand, and so on.
To this point, now we have achieved 80% accuracy with the present mannequin, and your entire knowledge pipeline is performing as per expectation. We now have additionally categorised suppliers into numerous segments primarily based on the supply timeline, resembling on-time, late or early.
The entire resolution has been examined and demoed to a peer group and printed over the Oracle market for the consumption of varied Oracle clients.
I’ve simply began Nice Studying’s PGP Synthetic Intelligence and Machine Studying Course. Nevertheless, in a really quick span, I’m able to brush up on my abilities in addition to get deeper insights from mentors, trade specialists, and numerous examine supplies.