Vinay Sankarapu & Arya.ai, An Aurionpro Company: Pioneering Explainable AI from India for Global Finance
Deep dive into Vinay Sankarapu, IIT Bombay alum & Forbes 30U30 AI founder. Explore Arya.ai's decade-long journey, tech (AryaXAI), BFSI focus & acquisition.
In the rapidly evolving landscape of Artificial Intelligence, few have been navigating its complexities longer or with more focused determination than Vinay Kumar Sankarapu. As the Founder and CEO of Arya.ai and AryaXAI, he stands at the forefront of making AI not just powerful, but also interpretable, safe, and aligned with human goals, especially in high-stakes industries like Banking, Financial Services, and Insurance (BFSI). This is the story of an IIT Bombay graduate turned AI pioneer, his decade-plus journey building one of India's first AI companies, and his vision for responsible AI.
Check out the video of the conversation here or read on for insights.
🎓 Early Days: IIT Bombay, Books, and an AI Spark
Vinay's journey began at the prestigious Indian Institute of Technology (IIT) Bombay, where he pursued a Dual Degree (B.Tech + M.Tech) in Mechanical Engineering, specializing in Computer Aided Manufacturing (CAM) from 2009 to 2014. Even during these early years, his diverse interests shone through. Remarkably, he became a published author of two novels – 'The Ark and the Aroma of Peril' (published via CreateSpace/Amazon in his first year) and 'Is She the One?' (published by Vitasta Publishing/Times Group Books in his second year) – all before turning 21!
Beyond writing, he gained early startup experience working as a Marketing Manager for an OpenEdex platform, one of India's early EdTech ventures, while still in his 2nd and 3rd year of college. This stint involved coursework development, digital marketing, and content creation, providing invaluable practical business understanding.
However, it was his Master's thesis in 2012/13 that truly ignited his passion for AI. He focused on "Using Convolutional Neural Networks (CNNs) in Turbine Engine Failure Prediction," diving deep into the complex mathematics of deep learning at a time when it was far from mainstream. This academic work laid the foundation for his future ventures.
💡 Genesis of Entrepreneurship: Tackling Information Overload with InvenZone (2013)
Vinay's research experience exposed him to a critical problem: the overwhelming explosion of knowledge. He realized that traditional research methods, involving months of literature review using basic keyword search tools like Google Scholar, were inefficient and often inadequate.
In 2013, more than 30 peer-reviewed papers were published every minute. Knowledge overload is only going to get worse. The current tools are very, very redundant.
Driven to solve this "knowledge suffocation," Vinay co-founded his first AI-focused venture in April 2013, operating out of an apartment near IIT Bombay. The vision was ambitious: create a platform and an AI research assistant for STEM researchers. With a seed round raised in early 2014 from around 20 angel investors (a feat in India's nascent VC ecosystem back then), a small team of 7 recent graduates embarked on this challenge.
We scraped the entire internet and indexed close to 34, 35 million research papers... We started to build the AI assistant for STEM researchers... We were using C++, plus plus, or there was an NVIDIA framework called Digits... we bought a 1080 [NVIDIA GPU] at that point in time, using it for deep learning. Probably we were the only guys at that time in India who was using deep learning on one zero eight zero.
InvenZone developed contextual search far superior to existing tools and experimented with predicting future research trends. However, they hit limitations. The AI assistant's conversational abilities weren't "human enough," memory management was an issue, and Natural Language Generation (NLG) technology wasn't mature. Commercial viability for a pure research tool also proved challenging.
🚀 Founding Arya.ai (2013) & The Pivot to Enterprise AI
Recognizing these hurdles, Vinay, along with his co-founder, made a crucial pivot in October 2013, founding Arya.ai. They repurposed the core deep learning technology developed at InvenZone, shifting focus from academic research to enterprise applications.
Initially conceived as a deep learning workbench for enterprises, Arya.ai worked with Fortune 500 companies across India and Europe. They soon found a strong product-market fit within the Banking, Financial Services, and Insurance (BFSI) sector. This highly regulated, data-sensitive industry desperately needed AI's power but struggled with deploying complex models that were explainable, auditable, and safe.
In 2015-16, we pivoted from the idea of building an assistant to a tool. Let's verticalize the platform instead of building a horizontal platform. Let's create more basis on that because you have tools, models, solutions, guardrails, and risk rails on top of it.
This vertical focus on BFSI became Arya.ai's key strength and differentiator.
🛠️ Arya.ai's Technology & Product Suite: An OS for Financial AI
Arya.ai evolved into what Vinay describes as an "Operating System for FSIs" designed to expedite the deployment of complex yet responsible AI. Their goal: solve the critical AI acceptability and scalability problem for banks and insurers.
Their platform comprises several key components:
Arya APIs: A collection of pre-trained AI models tailored for common BFSI tasks (e.g., customer onboarding, underwriting, fraud detection). Vinay mentioned Arya.ai processes the highest number of customer onboarding APIs for General Insurers in India.
Libra: A repository of task-specific, state-of-the-art deep learning models that clients can fine-tune on their own data.
AryaXAI (ML Observability & Explainability): This is a cornerstone of their offering, addressing the 'black box' problem. It plugs into ML workflows to deliver explainability, auditability, and safety monitoring for AI models. Key features include innovative techniques like DLBacktrace (a proprietary XAI algorithm co-patented by Vinay), alongside standard methods like SHAP and LIME, plus tools for monitoring model bias and data drift.
Explainability is the most important thing in financial, ah, highly regulated industries. If I'm a doctor and I tell you you should have cardiac surgery, but I can't tell you why, then it's a problem, right?
He further elaborated on a direct experience.
We built a deep learning model and tried to deploy it in a financial institution. We were rejected by a regulator and the risk manager because it was not explainable. No matter how accurate the model is, you couldn't understand what's happening inside it.
🌍 Vision & Impact: Democratizing Responsible AI
Vinay Sankarapu's overarching vision is clear: "to democratize complex 'Al' for highly sensitive, high-risk, and mission-critical use cases but make it interpretable, safe, and aligned to the user's goals."
He is a vocal advocate for responsible AI development and regulation, believing it's essential to prevent misuse and ensure ethical applications. He distinguishes between regulation (the rules) and governance (the allocation of responsibilities to comply with those rules), highlighting the complexity of implementing effective AI governance.
Arya.ai's focus on explainability and alignment directly supports this vision. The launch of AryaXAI as a distinct offering, potentially applicable beyond BFSI to sectors like healthcare and defense, underscores this commitment. To further this research, Arya.ai recently established a new research lab in Paris, aiming to become a leading center for AI interpretability and alignment.
We ensure these are interpretable and aligned so that the end user gets the advantage of the solution. By doing interpretability, can we do alignment intrinsically inside the model?
💰 Business Growth, Metrics & Acquisition
Arya.ai has demonstrated significant traction and business maturity. Vinay shared some key metrics during our podcast discussion:
Profitability: The core Arya.ai business has been EBITDA positive for the past three years.
Growth: The company experienced 3x growth over those three years.
Revenue: While not public, Vinay confirmed the revenue was broadly in the ₹50 Crore to ₹100 Crore ARR (Annual Recurring Revenue) range, with expectations to surpass that significantly in the coming year.
In a major milestone, Arya.ai was acquired by Aurionpro Solutions (a publicly listed Indian enterprise product company focused on BFSI) recently.
From that point of view, acquisition is a good route. Aurionpro, our acquisition partner, already has massive distribution access in the financial sector, which means we don't need to do the selling—we simply upscale the product.
This acquisition provides Arya.ai with greater firepower and distribution reach, allowing Vinay and his team to focus more intensely on product development and the cutting-edge research happening within AryaXAI, without the immediate pressures of independent scaling.
🏆 Recognition & Leadership
Vinay Sankarapu's contributions have earned him significant recognition, including being named in Forbes Asia's 30 Under 30 list for Consumer Technology. Arya.ai itself has received accolades like the "AI Game Changer" award from Nasscom and was named among the "Top 61 global companies" using AI in financial services by CB Insights. The company participated in prestigious accelerators run by Swiss Re, Société Générale, Microsoft, and Oracle, and was an early member of Intel's AI Builders and Nvidia's Inception programs.
🔮 The Future: Explainability, Alignment, and Beyond
Vinay Sankarapu's journey with Arya.ai exemplifies the evolution of AI itself – from ambitious academic concepts to practical enterprise tools, and now grappling with the critical challenges of trust, transparency, and alignment. With the backing of Aurionpro and a dedicated research focus through AryaXAI, Vinay is poised to continue shaping the future of responsible AI, ensuring that as machines become more intelligent, they also become more understandable and beneficial to humanity. His work underscores a crucial truth: the future of AI isn't just about capability; it's about building trust.
Listen now!
Other ways to listen:
Your Feedback matters
As always, I’d love to hear your thoughts! Whether it's about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.
Until next time,
Your Host,
Akshay Datt