COMPANY
One conviction,
turned into a stack.
AI must prove its outputs, not assert them. Everything Velkron builds follows from that sentence.
The last three years of AI were about generation: making machines produce text, decisions and advice at scale. The next three are about accountability. Regulated industries cannot run on output they cannot audit, and regulators have stopped accepting assurances in place of evidence.
Velkron exists to build the audit layer. Not another copilot, not a dashboard: the piece of infrastructure that sits between an AI system and the outside world and can say, with citations and a sealed receipt, what happened. Deterministic checks are the wedge. Receipts are the product. Cryptographic proof of inference is the destination.
The plan is deliberately sequential: prove it in UK financial promotions, port the engine to the Gulf rulebooks from Abu Dhabi, then let receipts and proofs extend to every regulated vertical.
Satyawan Singh
ML ENGINEER · SOLO FOUNDER
Every entry in this build log was shipped by one person in the last 18 months. The pattern is the point: every project makes machine output provable.
4 open-source verification libraries
calibrated abstention, output grounding, verifier-gated search, agent consensus
119M-parameter speech model
trained from scratch for UK clinical audio
MSc dissertation, University of Leicester
trust scoring for clinical AI
140-crate Rust blockchain
25,000+ passing tests, consensus verification primitives
RegTech retrieval engine
provenance-gated citations for FCA-regulated firms
Taking on a small group of design partners in regulated finance.
Wherever you're regulated: if your firm publishes financial promotions, or is putting AI anywhere a regulator can see, write to us. We reply to every request within two working days.