Caventia is the work of one founder with a specific track record and a specific argument.
Founder & CEO
Ashish K. Saxena
Amazon FinTech alum. Bestselling AI ethics author. IJSR peer reviewer. Two hundred and twenty-six citations across the literature.
- Amazon FinTech: 40% fraud reduction and 75% processing-error reduction at scale; 20% false-positive reduction
- Hospital management AI research: 30% wait-time reduction and 20% equipment utilization gain
- Author: "Society and the Machine" (2024 London Book Festival first place) and "The Ethics of Artificial Intelligence" - Amazon bestsellers in the US and UK
- h-index 8 on Google Scholar; 226 total citations spanning fraud detection, healthcare AI, AI policy and machine learning
- Peer reviewer at the International Journal of Science and Research (IJSR); 42 papers reviewed
- IEEE TEMSCON ASPAC and IEEE ISTAS 2024 contributor
- 2024 "Best Technical Researcher of AI", Business Innovation Awards · Marquis Who's Who
- Founder, MindBytesAI. Creator, ImpactLens AI. Fifty-plus AI professionals mentored.
Ashish K. Saxena is the founder of Caventia. He has fifteen years of AI deployment experience, including financial fraud detection at Amazon (40% fraud reduction, 75% processing-error reduction) and AI-driven hospital resource allocation research (30% wait-time reduction, 20% equipment utilization gain).
He is the author of Society and the Machine (winner, 2024 London Book Festival; second place Non-Fiction Education at the 2024 PenCraft Book Awards) and The Ethics of Artificial Intelligence. He serves as a reviewer for the International Journal of Science and Research (IJSR), has peer-reviewed 42 research papers and received the 2024 “Best Technical Researcher of AI” award at the Business Innovation Awards.
His published research spans machine learning fraud detection (Emerging Trends in Machine Intelligence and Big Data, 2019), healthcare AI (an LSTM resource-allocation paper in the International Journal of Applied Health Care Analytics, 2022) and AI policy. His paper on AI integration (“Decoding Socioeconomic Influence on AI Integration and Trust in the U.S.”) was accepted at the 2024 IEEE TEMSCON ASPAC. His work on bias measurement in AI-generated content was presented at the 2024 IEEE International Symposium on Technology and Society (ISTAS). Google Scholar reports an h-index of eight with two hundred and twenty-six citations across the literature.
At Amazon FinTech I watched smart engineers ship models that moved billions of dollars in payments and I watched the model risk function struggle to keep up. We were good at the engineering. We were not great at the documentation an examiner expects under SR 11-7. The gap kept widening.
Then generative AI arrived. Suddenly every business unit wanted an agent. Suddenly every agent was making decisions that touched fair-lending, fraud, KYC and clinical care. The horizontal AI governance vendors that sprang up in 2023 and 2024 (Credo, Fiddler, Arthur) were doing useful work, but none of them were shipping the artifact a Federal Reserve examiner asks for. The language did not match. The schema did not match. The mental model did not match.
Caventia exists because the next decade of AI inside regulated industries needs a platform whose first principle is not “make AI safe” but “make AI legible to the specific regulator who is going to read it.” That is a vertical problem. SR 11-7 is not the same as the FDA's 510(k). 510(k) is not the same as ECOA. The artifact your OCC examiner expects is not the artifact your IRB expects. We built the platform around the artifact, not the other way around.
The bet is simple: in the regulated half of the AI market, the winning platform will be the one whose evidence ledger is examiner-ready by construction. Caventia is that ledger.