As a part of the Uber team in 2015, Kunal Singh had seen how the mobility giant had started experimenting with AI (artificial intelligence) and ML (machine learning) to streamline work across a variety of functions: churn modelling, pricing, customer discounts, and marketplace matching, among others.
However, despite AI/ML systems being promised as a gamechanger for enterprises, Kunal realised that even until 2020, a majority of companies were still struggling to implement and build on emerging tech.
It was then that Kunal started in Ahmedabad in 2020.
In partnership with a Mulesoft-based firm called Eshia Solutions, Discite began its journey as a full-stack AI/ML and data solutions services startup. Mulesoft is a software company based in California.
“Our mission is to democratise artificial intelligence for enterprises of all sizes, globally. We help our client companies either start or progress on their AI strategy and journey through custom-made solutions, with our major focus areas being Computer Vision, Natural Language Processing, Deep Learning, MLOps, Data Engineering and all forms of analytics,” adds Kunal.
The company recently raised an undisclosed round of funding led by US-based IT firm DynPro. The team has since moved its base to Bengaluru.
How does it work?
The startup’s prime focus is on building AI/ML solutions that add value to each organisation based on their need.
Kunal explains, “Before we write even a single line of code, we have a lot of sessions with various stakeholders and end-users within their organisation to understand their processes and their pain points. Once we build out the solution/product, we work closely with the client’s team till the time the desired business impact takes place. We also perform the role of thought partners, and explore other areas within the client’s organisation where bringing in AI/ML can give them a significant boost in highly competitive markets.”
The biggest challenge for the team has been hiring and retaining talent. Data Scientists, ML Engineers and Data Engineers are all in great demand globally, whereas relatively fewer of these specialists are actually available for hiring.
What sets Discite apart are its engagement models:
Project-based: The team builds a particular solution and then transfer/integrate it into the client’s ecosystem. Discite charges on the basis of its time and resources that go into building and maintenance of this solution.
Resource-based: Sometimes when a client needs one or more Data Scientist/ML Engineer/Data Engineer to work on various things within their org, the team also deploys their employees on a flexible basis to work part-time or dedicatedly with them.
Revenue-sharing based: This is a new model that the team is working on. Descite will be building a product at a much-discounted cost in partnership with an enterprise or startup founders, in return for a percent of the revenue (for a fixed period of time) that the product generates.
Team and market
The startup is currently run by a 30-member team. The first hire was Moulindu Sen, who is now Co-founder and CTO.
“In his years with Deloitte’s Applied AI team, Moulindu was making a similar observation as me – that companies of all sizes throughout the world want to implement more AI but severely lack the guidance and the people to actually do it. So when we talked about the possibility of him working Discite, he was instantly interested,” the founder says
Tanya Mehta, who joined as the head of growth, is an old friend of Kunal’s.
According to IDC the India AI software market will grow from its $2.76 billion in 2020 to $6.35 billion in 2025 with an 18 percent CAGR. Some of the key players in the segment include Neptune AI, data iku, Reckon and others.
Revenue and future
The team refused to share the exact amount it charges. However, Kunal says, “In terms of the business model, we’re similar to any tech services-based company. We take up projects/resource-deployments and charge our clients based on the time spent by our Data Scientists, ML Engineers, Data Engineers or Analysts on the project. The margins are generally in the range of 20-30%.”
The fresh funds from DynPro was also a strategic call for the team as the IT solutions company has an impressive clientele, including Airbnb, Block, Lucid Motors, Ebay, and Sony, among others
While the team declined to name the clients or the sectors, due to legal guidelines. Kunal states, “Within our first year, we’ve already worked with more than 10 enterprises and startups, across the US, Middle East, Australia and India.”
The team is on a hiring spree and plans to reach a team size of over 60 members and an ARR of $1 million by the end of 2022.
“We are also in the process of building ML-based B2B SaaS products of our own. We’re also planning to work on more upcoming areas within AI/ML; like Self Supervised learning, Deep reinforcement learning, Adversarial learning, GAN (general adversarial network) based data augmentation, FSL (few shot learning), among others,” says Kunal.
