Pritesh Tiwari, Director of Data Science, VYugma and Founder & Data Scientist, Data Science Wizards
Sandeep Khuperkar, Founder, VYugma and Co-Founder & CEO, Data Science Wizards
1. Building Graph
Accelerated AI
Solutions at Scale
BY
Sandeep Khuperkar,
Founder, VYugma and Co-
Founder & CEO, Data
Science Wizards
Pritesh Tiwari, Director of
Data Science, VYugma and
Founder & Lead Data
Scientist, Data Science
Wizards
1
WHERE OPEN SOURCE MEETS INNOVATION
VYUGMA TECH SOLUTIONS PVT LTD. (CONFIDENTIAL)
2. AI Trends/ AI Initiatives
+14% Global GDP in 2030
$15.7tr
Potential contribution to the global economy by 2030 from AI
Top Sectors and potential impact on scale of 1-5
● Healthcare (3.7)
● Automotive (3.7)
● Financial Services (3.3)
● Transportation and Logistics (3.2)
● Technology (3.1)
● Retail (3.0)
● Energy (2.2)
● Manufacturing (2.2)
Source: Gartner
VYUGMA TECH SOLUTIONS PVT LTD. (CONFIDENTIAL)
Source https://www.pwc.com/gx/en/issues/data-and-
analytics/publications/artificial-intelligence-study.html
3. Challenges we are witnessing
● Lack of a 360 degree view of data which can help in data management and data drift (change over time)
● It is difficult to manage data which has no schema and this is a common case when dealing with large
amount of raw data further used for training AI ML models.
● Difficulty in finding association between different types of data available.
● New Use Cases needs to build from scratch even though we have the data available.
● Time taken for data discovery is high when the associations are not known.
● Require help of expert Data Scientist with Domain Expertise to dig deeper into data and find common
points.
● Proactive identification of new insights from the data already available.
VYUGMA TECH SOLUTIONS PVT LTD. (CONFIDENTIAL)
4. How Graph Technology (neo4j) is helping
to enhance the effectiveness of AI Solutions
● Helping create transparent and unified access to diverse and heterogeneous data sources
● To integrating various sources into a single view and easily accessible database.
● The benefit and value of a graph database is that the data is modeled and presented to the user as a
graph including unstructured data.
● In graph models, the separation between schema and data is less explicit as compared to relational
model thus help handling highly interconnected data.
● With the help of associations new use cases can be built faster and much efficiently.
VYUGMA TECH SOLUTIONS PVT LTD. (CONFIDENTIAL)
5. Source : Article A Knowledge
Graph-Based Data Integration
Framework Applied to Battery
Data Management Tahir Emre
Kalaycı * , Bor Bricelj, Marko Lah,
Franz Pichler, Matthias K.
Scharrer and Jelena Rubeša-Zrim
https://www.mdpi.com/2071-
1050/13/3/1583
5
VYUGMA TECH SOLUTIONS PVT LTD. (CONFIDENTIAL)
9. Solution 3- UnifyAI
VYUGMA TECH SOLUTIONS PVT LTD. (CONFIDENTIAL)
Three Key Building Blocks of
UnifyAI Platform
1. Data Pipeline
2. Machine Learning & AI
3. UnifyAI Orchestrator
10. Graph Accelerating AI Solutions and Enterprise Adoption
● Along with managing large volumes of data, today’s business decision maker also need to generate
insights from existing data, and graph database efficiently helps to discover and store the
relationship between data features.
● Companies using a graph database in conjunction with other relational and NoSQL database
management systems can enjoy sustainable competitive advantage by using real time data
relationships and making faster decisions.
● Handing data relationship are difficult with legacy systems like RDBMS because of their rigid schema
structure which makes it difficult to add different connections or adapt to new business requirements.
● Implementation of graph algorithms after identifying the relationship between data always help in
building efficient machine learning models.
VYUGMA TECH SOLUTIONS PVT LTD. (CONFIDENTIAL)