I work at the boundaries of software — where it meets hardware, and where information is incomplete, ambiguous, or uncertain.
By day, I’m a technical consultant with experience in investment management, working end-to-end across investment management systems (eFront, DTCC CTM). Outside that, I research and build around one question:
How should a program define uncertainty, and represent it accordingly?
I study how existing systems and models represent incomplete knowledge, weigh evidence, and turn it into decisions. That covers programming languages and formal methods, information and decision theory, and the constraints that make hard problems tractable.
Papers, technical reports, and working notes are on Epistemia.
I test ideas by building them. Lana is my answer to that question in code: a language, compiler, and virtual machine where uncertainty and decisions are ordinary values. It currently powers Living Canvas.
To see why Lana exists, start with Can uncertainty be programmable?
I also built an MCP server on top of Qualcomm's qai-hub SDK, so AI assistants can compile, profile, and run models on real devices through Qualcomm AI Hub.
- Start with a concrete question, break down its semantics, and write down the precise claim before building anything.
- Make it checkable: pinned versions, experiments you can re-run, and numbers from real runs.
- BS Computer Science, University of Wisconsin–Madison
- MS Financial Management (Investment Analysis), Boston University
For research, collaboration, business inquiries, or general correspondence, email vijayneelakantan@outlook.com.
Based in Boston.



