[100 Emerging Women Leaders] From hospitality to building LLM, how Preksha Kaparwan realised her passion for tech

Upon finishing school, Preksha Kaparwan’s parents gave her two career options: study engineering or enrol in a course of her choice in a renowned college in Delhi. 

While her heart was in tech, Kaparwan wanted to avoid the competition that came with an engineering course. She opted instead to study hospitality at the Institute of Hotel Management, Pusa, Delhi. However, she worked for a couple of years in the hospitality sector as a sous chef at The Imperial before quitting that job.

She had switched to teaching consumer behaviour and financial services at Pearl Academy, Chandigarh, when life gave her another chance to pursue her dream. She met Saurabh Moody on a chance encounter; he was creating cashless systems of transactions for the hospitality industry. 

“Those were the days you did not even have QR codes in Paytm,” Kaparwan tells HerStory. “And since I had a strong network with hoteliers, we started working together.”

During the early years of collaboration, the owner of a multichain restaurant approached her to create a food delivery website and Kaparwan found it the perfect time to get hands-on with tech. “I got on YouTube tutorials and decided to learn coding. In a week I had the first version ready,” she recalls. 

She and Moody have together built three startups: Realbox, a real-time data analytics company; Cashless, a payment technology platform; and most recently, Super.ai, an enterprise AI startup for enterprises that helps them generate insights from structured and unstructured data using private Large Language Models (LLMs) on-premises.

Simplifying human-machine communication

In their previous venture Realbox, Kaparwan and Moody had created a universal database connector but realised that “every product platform needs a different kind of connector for a different kind of database,” says Kaparwan.

When they found themselves working as an extension to the analytics team for enterprises, they chose to exit their ventures. “These enterprises kept coming up with requests for specific questions to be added or data to be pulled out. We weren’t supposed to do that because we were building a product here,” she explains.

This fuelled the idea of natural language interaction with data.

Under Super.ai, Kaparwan has created a platform similar to ChatGPT; an AI tool for enterprise data that allows users to ask questions in natural language, generate comprehensive stories on their data, and integrate with existing Business Intelligence (BI) tools such as Tableau and Power BI. These solutions are designed to extract, transform, and present data to support data analysis, trend identification, and strategic decision-making within the organisation. 

“Super.ai aims to unify business knowledge insights with data to create context and relevance across the ecosystem,” Kaparwan says.

The duo spent six years developing the technology, creating their own native Large Language Model (LLM) as opposed to using an Application Programming Interface (API) from Open AI or anywhere else. 

“We train an enterprise AI by collating insights from unstructured and structured data in a private LLM environment that can be accessed by everyone at any skill level with a simple conversation,” she explains.

Super.ai, which last raised $1.8 million in 2022, is currently working with 14 companies, including DAMAC Group, Incorta, Procter & Gamble, MiQ, Accenture, and some pharma companies.

Gender biases

Although India’s IT sector has the highest female participation rate of 30% (according to the Business Responsibility and Sustainability Reporting disclosures by the CFA institute), Kaparwan believes only a few women techpreneurs are able to break the glass ceiling despite being highly skilled. 

Despite many opportunities at the business level in tech, gender bias still abounds.

“Once you are within the inner tech circle, gender lines are blurred and talent is prioritised. But not many women get to this space so easily because they are not taken seriously. They have to constantly prove themselves,” says Kaparwan.

However, when it comes to the core technical team, the field has in fact been levelled for women techies, she believes.

“I often see that women are affected by the initial struggles and the assumption that they won’t be accepted in the hardcore tech space. This affects what they bring to the table too,” says Kaparwan. 

“If you have done your graduation and the initial preparation, you are at the right time and place to grow in this field. You realise that beyond gender binaries, innovative minds matter, given the grave talent shortage in the industry now,” she adds.

Kaparwan and her team are now working towards creating Enterprise General Intelligence, a live AI brain for businesses that can diagnose and suggest actions based on historic market situations. “We are already executing parts of it but a lot is still underway,” she says.