Our Origin

DNAi Systems was founded by two physicians who saw the same problem from both sides of the exam room: AI is transforming healthcare and professional life, but you can’t trust what you can’t verify.

“We’re not AI researchers who pivoted to healthcare. We’re a psychiatrist and a pediatrician who built the AI infrastructure we wished existed — for our patients first, and then for everyone who needs an AI that cannot lie, must say when it doesn’t know, and always puts the user first.”

That conviction led to a fundamentally different architecture. Over 20,000 conversations — each one testing, breaking, and rebuilding the rules — crystallized into a set of core principles. Those principles are now enforced by code. The result is a cognitive system with over 100 million knowledge nodes, each stored as a structured “knowledge atom” with a Name (what it is), a Form (how it manifests), and a Dharma (what it does and when to use it).

Every atom is cryptographically hashed and sealed. No single fact can vouch for itself — every recommendation requires a second, independent atom to validate it, the way a surgeon cannot operate without an anesthesiologist confirming it’s safe to proceed. The model becomes a voicebox. The architecture does the reasoning.

The Architecture at a Glance

100M+
Knowledge Nodes
346+
Indexed Collections
3
Granted/Filed Patents
0
Data Sold. Ever.

Each knowledge atom — called a Cognitive Inference Unit (CIU) — carries a cryptographic hash that chains it to its parent, a semantic vector embedding for intelligent retrieval, and a confidence-falsifiability score that determines whether it should be trusted. Modify any unit and every descendant hash breaks. This is a patent-pending architecture (US 19/290,471, allowed; additional provisionals filed) designed for domains where trust is not optional.

Principles

Fiduciary by Design

Our systems act in the user’s interest. An architectural constraint: harming the user is unreachable through any inference path.

Transparency

Every AI response links to its evidence. If it can’t be sourced, it doesn’t get said. The entire reasoning chain is auditable.

Epistemic Humility

When confidence falls below threshold, the system says “I don’t know.” Built into the architecture. Not a polite fallback — a hard constraint.

Privacy by Architecture

Tenant isolation, zero data retention from AI providers, encrypted in transit. Your data stays yours. Always.

Founders

Dr. Deepan Singh

Deepan Singh, MD, FAPA

Co-Founder & CEO

Double Board-certified psychiatrist specializing in child, adolescent, and adult care. Department of Psychiatry at Maimonides Medical Center, Brooklyn, NY. Past Vice-Chair of Ambulatory Services and Chief of Research & Grants. Key opinion leader in Prader-Willi Syndrome research. Named inventor on the DNAi patent portfolio. Designed the cognitive architecture that powers Asha and Harley.

Dr. Paridhi Anand

Paridhi Anand, MD

Co-Founder

Double Board-certified pediatrician specializing in behavioral health, collaborative care, and obesity medicine. Over 15 years of clinical, academic, and research leadership experience at NYC Health + Hospitals. Co-creator of Asha, and Harley. Expertise in preventive medicine and the design of AI-native clinical systems that serve real patients in real time.

Team

Edwin Servellon

Product Lead, Harley AI

Founder of Servellon Fitness. Edwin ran his entire training business on Trainerize before becoming the first trainer to migrate to Harley AI. Years of building workouts, filming demonstrations, and managing a growing client base by hand — his daily workflows, pain points, and product vision shaped every feature Harley has.

Products

Company

Incorporated
Delaware, USA
Headquarters
Roslyn, New York
Website