Accountability stays with a person
A named person signs and stays accountable for the decision. In Serlyx, the underwriter decides. Models assist, explain, and narrate under that person's control.
Company
VedAGI builds human-aligned AI for high-stakes decisions, in two products. Pramana (working name), our lead product, is the lock on the figures in market access and evidence packs. Serlyx helps underwriters assess AI risk. In both, AI supports the person and a named human stays accountable. Everything we build is local and offline-first by default.
Mission
Vision
Designed for enhancing human lives and work for the betterment of society.
Human-aligned, for us, means the person who carries a high-stakes decision stays accountable. AI strengthens that person's judgment and never replaces it.
Why we exist
AI is moving fast into life sciences. Teams now use large language models (LLMs) to build models, summarize data, and draft packs. An LLM can produce a figure that looks right but cannot be traced to a source. That is fine for drafting an email. It is not fine for a budget-impact table, a dossier, or a submission that a payer or health technology assessment (HTA) reviewer will use to decide coverage. The problem is architecture, not model size. When the same system that writes the prose also produces the numbers, nobody can tell what was computed and what was guessed. VedAGI separates the two. AI may plan the work, explain, and narrate. Code computes the numbers from source. A named human signs and stays accountable.
How we build
A named person signs and stays accountable for the decision. In Serlyx, the underwriter decides. Models assist, explain, and narrate under that person's control.
Use the model path your policies allow. In Pramana (working name), models may plan the work and propose a formula, but code computes the value.
In Pramana (working name), each figure is computed in code, checked against source, and recorded with its sources, its formula, and who signed it.
In both products, an estimate is labelled as an estimate. In Pramana (working name), when evidence is thin, the output is labelled ESTIMATE, so it never passes as a computed figure.
Your data stays inside your controls. VedAGI is not your system of record.
Deploy where the work happens, including environments that cannot depend on a live connection.
Products
Pramana (working name): the lock on a number before it leaves the building. AI can draft the pack. It cannot be the source of the figures.
Serlyx: helps underwriters assess AI risk and E&O (errors and omissions) exposure for companies that build or use AI. A human underwriter makes the decision.
Leadership
Founder & CEO
Rajan founded VedAGI to build AI that the people accountable for high-stakes decisions can actually check.
He has spent more than twenty years building enterprises and running the technology inside them. He founded an executive consulting firm and grew it over a decade to more than $10 million in annual revenue, leading IT and MedTech growth strategy and digital transformation for Fortune 500 clients, with cybersecurity, regulatory compliance, and data management built into the work. Those programs delivered annual cost savings of around 30 percent for clients.
He also serves as President and CIO of a medical instrument manufacturer that supplies hospitals and clinics in North America and internationally.
He holds an MSc from the University of Liverpool and BRMP, COBIT 5, and ITIL V3 credentials. His full work history is on LinkedIn.
Join us
If you want to build AI that can be checked, or partner with us, get in touch.
We work with a few market access, HEOR, and evidence teams at a time, because doing this properly means getting into the real work.
If you are about to send a table you cannot recompute, talk to us.
VedAGI. Intelligence for Humanity.