I build AI agents for enterprise finance.
I’m most interested in what happens after AI starts doing the work. Not just whether a model can complete a task, but what happens when we realise that some of the work, steps, approvals, and handoffs we built around that task may no longer be necessary.
I’m on the founding team of cleartax.ai, Clear’s agentic AI platform for the CFO office. My work spans building the product, working with enterprises on real-world POCs, and figuring out what AI agents can reliably do inside regulated finance workflows.
Before AI, I spent years building the infrastructure that connects enterprise finance systems to the tax and compliance workflows around them.
What I’ve built
I’ve been at Clear since 2018. I started by building the ERP integrations function from scratch, including FTP Magnet, a multi-tenant middleware platform, the Tally Add-On, and ClearTax for Developers, our consolidated API platform and documentation.
I also led Clear’s expansion into the GCC, working across pre-sales, implementation, and delivery.
That function eventually grew into an 18-engineer organisation delivering hundreds of enterprise integrations every year across SAP, NetSuite, Oracle Fusion, and Microsoft Dynamics, supporting GST, e-invoicing, and TDS compliance.
The most rewarding part of that journey came later, when I started applying AI to the work itself.
I co-architected our AI Integration Agent on Claude Code, where engineers and AI agents work together on a shared interface. It reduced implementation time by roughly 80%.
I also built Pulse, an LLM-powered system that replaced traditional standups with a live view of project health, turning Slack updates into signals across hundreds of concurrent projects.
These experiences changed how I think about enterprise AI.
How I got here
My career started in finance.
I studied commerce at Canara College, Mangalore University, and began my career in audit at Manohar Chowdhry & Associates. I worked across bank audits, SEZs, oil and gas, tax audits, and multi-currency accounting, using systems such as Tally, SAP, and Finacle along the way.
I kept building small tools to analyse the data I was working with. Eventually I realised I enjoyed building the tools more than using them for audits.
In 2015, I founded OmniMonk Solutions, a web technology company. We built interactive EPUB3 books for the World Konkani Centre, including narrated audio and synchronised text highlighting. One of those projects was a runner-up for Digital Book of the Year at the Publishing Next awards.
In 2017, I joined Vicit InfoTech as a founding member and lead developer, where we built IconicBot, a multilingual AI chatbot for celebrity fan engagement. The team was selected for Canada’s NEXT AI program.
Then I joined Clear.
Looking back, the thread is fairly simple: I’ve spent my career working where finance meets software, and increasingly where software meets AI.
What I write about
I write mostly about AI in the enterprise, especially what happens when AI changes processes rather than simply automating tasks.
I also write about technology and society, philosophy, psychology, and the occasional musing on life in Bengaluru.
Outside work
I read a lot of non-fiction, listen to classical Indian music, and spend time thinking about evolution, technology, and why people and organisations behave the way they do.
I also practice yoga and try to make it to the gym more regularly than I do.
I’m based in Bengaluru, India.