Kumar Aditya.
Architecting mathematical trust for enterprise AI.
Born and based in Ara, Bihar. My work focuses on deterministic runtime execution, zero-trust client-side evidence anchoring, and multi-agent coordination frameworks. I founded Arkashri Private Limited to establish institutional-grade software that survives adversarial and regulatory scrutiny.
Key Achievements & Publications
IEEE-Published Research Author
Authored and published peer-reviewed research in IEEE on client-side cryptographic commitments across heterogeneous blockchains, guaranteeing audit integrity under compromised servers.
Founder & CEO · Arkashri Private Limited
Founded Arkashri to develop 8 production-grade enterprise AI operating systems covering multi-agent audit compliance, non-profit governance, education, and algorithmic risk.
Autonomous AI & Agent Systems Architect
Engineered autonomous AI agent orchestration libraries, persistent memory cognitive graphs (JAY), and deterministic LLM execution replay pipelines.
Enterprise Deployments for Indian Institutions
Built and deployed mission-critical systems: NMEC ERP managing NGO statutory compliance (12A/80G), and NexSchool ERP modernizing K-12 institutional operations.
GitHub Repositories (@KumarAditya1729)
From Ara, Bihar to Global Architecture
A small town, a large curiosity
Born and raised in Ara, Bihar — where the state moves slowly and imagination has to compensate.
First systems
Wrote software for people I could see. Realized the interesting problems were never in the code — they were in what the code was allowed to promise.
AI systems
Moved from applications to foundations — multi-agent systems, retrieval, model orchestration. Became obsessed with trust in generated outputs.
EVIDEX
Published in IEEE on zero-trust client-side anchoring. Understood cryptographic evidence would become the load-bearing wall of enterprise AI.
Arkashri
Founded Arkashri Private Limited. Not a startup — an institution. Mandate: trustworthy AI infrastructure for the enterprises of Bharat.
The ecosystem
TrustOS, JAY, EVIDEX, NMEC, ORNEXA, NexSchool, Kuber, ResolveX — eight products under one thesis.
Four questions I owe an honest answer to.
Why do I build?
Because trust, once lost, is not restored by law — only by proof. I build so that proof is cheaper than doubt.
What am I building?
A research-first institution whose products behave like public infrastructure — auditable, boring in the best sense, and quietly correct.
What principles never change?
Evidence over marketing. Reversibility over speed. Institutions over founders. Silence over spectacle.
What kind of company am I trying to leave behind?
One that outlives me — measured in decades, not funding rounds; judged by what it made unnecessary, not what it made popular.
Six rules I write for myself, every quarter.
- 01
Ship the paper before the pitch. If it does not survive review, it will not survive scale.
- 02
Assume the auditor is smarter than the engineer. Build accordingly.
- 03
Every product must earn its place in the ecosystem by proof, not by ambition.
- 04
The interface is a contract. Do not break it lightly; do not decorate it at all.
- 05
Software is a promise. A promise that cannot be verified is a lie.
- 06
Do the boring thing beautifully. Return to it in ten years and it should still hold.