I’m an engineer, researcher and founder based in Pune, India.
My background spans electronics and race car engineering, a Masters in Technology Systems Management from Stony Brook University, and five years as a Senior DevOps Consultant across Azure cloud infrastructure in New York — building systems at enterprise scale across Fortune 500 environments.
In 2023 I stepped away from industry to pursue something I couldn’t find inside it — genuine first-principles research. That led to a geometric theory of knowledge emergence in AI, two patents, and the founding of Crucible Science.
I work at the intersection of information geometry, epistemology and AI architecture.
Work
Research
- Theory of Knowledge Attention Fields — KAFT Framework (arXiv preprint)
- Recursive Geometry Mutation Knowledge Dynamics via Covariant Fisher-Rao Attention (Provisional Patent, Nov 2025)
- Dynamic AR Content Optimization using Model Context Protocol (Complete Patent, July 2025) (https://github.com/ashwinpalve/ar-bookart)
Code
Publications
- Race Car Monitor and Control using Telemetry — IJSETR (2014) (http://www.ijsetr.com/user/issue.php?issue=ISSUE%207&volume=Volume3&page=3)
Projects
Crucible Science
Research intelligence lab.
Geometric AI for patent analysis, drug discovery, domain research.
Wonder
Geometric lens AI engine. Alpha access via Crucible Science.
Book AR
Dynamic AR content pipeline via MCP coordination. (In development)
Now
Updated February 2026
Finalising KAFT paper for arXiv submission. Building Crucible Science platform with Wonder engine alpha. Exploring geometric LLM training architecture as next experimental frontier.
Contact
If something here resonated — that geodesic is worth following.