Fireside Chat: Unveiling the vision- A Q&A session with Subhodip Pal on how Neo4j and Snowflake can be leveraged to modernize Higher Ed Advancement

Good Morning Mr. Pal, thanks for your time. Let's get straight to the point, Chat GPT and Generative AI, how is it going to disrupt the Higher Ed Advancement

It's Great for generative content where the cost of failure and going wrong is low and rectifiable

Such as?

Drafting Personalized Communication, Invitation, Emails & Newsletters, generating and embedding personalized Donation History or Engagement Levels.

These personalized interactions driven by hyper personalized stats will definitely act as a catalyst to increase engagement level and opens up channels of conversation.

It is a wonderful tool to enhance creativity and productivity for content generation for PR.

How about Chatbots?

Chatbots for automating Level 1 conversations are fine beyond that there are quite a few things that we need to be aware of for large-scale real-life scenarios like Bias, Dataset on which it's trained, how do you explain the content generated from complex mathematical models on a day-to-day basis and then fine tune it based on the feedback and business requirements. These are sophisticated engineering tasks.

Is that enough?

No, its not. LLMs doesn't solve all the problems in the world, it's just a branch of AI. There are multiple other tools and technologies that are available to us , which when leveraged efficiently can produce amazing outcomes.

Like?

Just like how we have seen an evolution in the speech end there is similar revolution happening on the vision side.

Interesting...

With these tools, let's imagine if we can capture the data around who is attending an event, who is interacting with whom - how long or how frequently, what are the various clusters from Images and video recordings through sophisticated modern tools like Snowflake. That will create an invaluable and rich network of data which when projected to a graph network can start generating valuable and actionable insights.

Do you have a visual to understand better?

Definitely, let's examine two frames of data from a demo from a Canada based company named, Prodco, projecting the movement of people in a conference.

If we can project the interactions from various frames, it can give us the ability to apply Graph Data Science and generate quite a bit of insight, like

? Communities - What are the various people who share a common interest or interact frequently

? Betweenness Centrality - How can we find an expert with rare dual specialty like Mathematics and Biology for Vaccine study

? Eigen Vector - The red dots demonstrate the relative importance of a person. Here the biggest red circle may be a famous AI scientist whose appointments are very hard to get. So, people try to establish connection with the people closer to him, smaller red dots, for an introduction

This is amazing, how do one start their Modernization Journey?

Good Question. First, we need to collect data around a constituent's journey, right from Application to Student to Alumni in a Constituent 360 Graph. Then with the help of 3rd party public and 1st party datasets keep updating the Graph with relevant information.

So now that we have the data around the constituent and we have a way to understand their affiliations, interests, hobbies, causes that they were/are interested in etc., the next step will be to track the donations and related transactions.

Once we start collecting these activities, then we can very easily classify (in terms of Time, Talent and Resource) which Constituent is interested in

-- Donation and for what cause and how much

-- Who are interested to volunteer?

-- Who are the Crowd puller and influencers?

So now we have a fully functional in-house social network?

Yes, and then once you have the foundation, you can start leveraging AI to personalize the donation suggestion and targeting based on the rich data sets available on the graph based on not only their past donations but also based on their interests like the clubs /activities they were part of during their college years.

Does it have the capability to foster collaboration and innovation?

Yes, AI driven Alumni Networking can quickly help to identify and filter through the candidate list for collaboration using algorithms like Page Ranks and Link Predictions.

We can also leverage something like Snowflake to train a base pretrained LLM with Institution specific knowledge base, white papers, research, awards, grants etc. and that can then assist in writing Grant Application, personalizing communications, finding others with similar domain of research etc.

What else Snowflake can do?

Snowflake can enable a wide range of use cases:

1. Vector Search - Match the image in the database to incoming guest and greet with a personalized greeting.

2. Timeseries Prediction (Cortex AI) - Based on the recently released turnkey ML solutions, organizers can now reliably predict the number of participants for an event

3. Streamlit APP - Event specific vertically integrated AI driven apps can be built. Snowflake Native App Framework and Marketplace has the potential to become

the B2B version of Apple App Store.

Switching topic, I have been hearing a lot about Snowflake Data Cloud being leveraged extensively in the BFSI sector, is there any potential application in the Higher Ed?

Absolutely, most major Universities have multiple Campuses which have their own IT systems and infrastructure. With the current setup, it's really difficult to accurately report various metrics to the State and Federal agencies. Securing multiple copies of sensitive PII data across multiple systems is also a daunting task.

On the contrary adopting Snowflake's Data Cloud for Private (Inter-Campus) and Public Data Sharing while centrally controlling the governance, masking and access without physical replication of data is a game changer. Here is one such implementation for an esteemed University with Four Campuses, each running on different Cloud Platforms.

Thank you, Mr. Subhodip Pal, for the wonderful insight into the shifting trends in the market and technology.

Subhodip Pal is a Director at a leading Management Consulting Firm with more than 18 years of experience in the Enterprise Data Management domain implementing and researching various tools, technologies in the AI, Big Data, Cloud, Data Governance, Master Data, Data Quality and Graph Data Science Algorithms. He holds various certifications in cutting edge products like Snowflake, Neo4j as well as certified in Graph Analytics for Big Data from UC San Diego, Machine Learning and Deep Learning from Microsoft.

COMTEX_447101075/2850/2024-01-31T08:07:08