Saram ConsultingA Day in the Life: How AI Agents Transform the CSV Engineer's 8-to-5
Hour by hour, a Computer System Validation engineer's workday before and after deploying AI agents — from document factory to validation orchestrator.
Read Article →Saram Insights
Thoughts and strategies on the intersection of Biotech, Artificial Intelligence, and Regulatory Compliance.
How Linear Agent's Architecture Solves the CSV Compliance Problem
Linear's five architectural decisions for building AI agents — boundaries, constrained tools, domain skills, contextual approval, and a custom harness — map directly to the regulatory requirements of Computer System Validation in life sciences.
Read Article →Why Your LLM Fact-Checker Is Lying to You (And the Architecture That Actually Works)
When Model B checks Model A from memory, you get correlated hallucination, not verification. Here's the evidence-grounded pipeline that cuts hallucinations 30-50%.
Read Article →MCP Authentication Decoded: How OAuth 2.1 Secures Remote AI Tools
The Model Context Protocol delegates security to OAuth 2.1, but the details matter. A protocol-level walkthrough of the discovery flow, token lifecycle, Enterprise-Managed Authorization, and the seven-layer defense model for regulated deployments.
Read Article →The Reference Architecture for AI Agents in Regulated Life Sciences
The 7-layer reference architecture that turns individual AI compliance components into one integrated, inspectable, FDA-defensible system — from identity to evals.
Read Article →