Simor

Data Infrastructure for Production AI

Practical writing on AI data engineering, feature stores, and the infrastructure choices that determine whether AI systems work in production.

From Single-User to Multi-User: The Ten Controls You Need Before You Scale
From Single-User to Multi-User: The Ten Controls You Need Before You Scale
26 Jun, 2026 | 11 Mins read

An AI application built for a single user has no tenancy concerns. The user is the user. There is no data isolation problem because there is only one data set. There is no cost attribution problem bec

Agent Guardrails: Containing What an Agent Can Do in Production
Agent Guardrails: Containing What an Agent Can Do in Production
25 Jun, 2026 | 09 Mins read

Input guardrails check whether a user prompt is safe. Output guardrails check whether a model response is appropriate. Agent guardrails check whether the actions an agent takes are within bounds. Thes

CI/CD for ML: MLflow vs Weights & Biases vs Neptune
CI/CD for ML: MLflow vs Weights & Biases vs Neptune
25 Jun, 2026 | 05 Mins read

Machine learning teams face a version control problem that Git does not solve. Git tracks code changes, but ML experiments change more than code — they change hyperparameters, datasets, model architec

Multi-Agent Failure Modes: What Breaks When Agents Call Agents
Multi-Agent Failure Modes: What Breaks When Agents Call Agents
24 Jun, 2026 | 10 Mins read

Single-agent systems have predictable failure modes. The agent calls a tool, the tool fails, the agent receives an error and decides what to do next. The failure is contained to the single agent's con

MCP in Production: Registry, Auth, and Permission Models
MCP in Production: Registry, Auth, and Permission Models
23 Jun, 2026 | 11 Mins read

The Model Context Protocol gives AI agents a standardized way to discover and invoke external tools. In development, MCP works well with a local server running on localhost and a handful of tools. The

When your AI vendor goes bankrupt — surviving platform lock-in
When your AI vendor goes bankrupt — surviving platform lock-in
23 Jun, 2026 | 05 Mins read

A healthcare analytics company received notice on a Tuesday afternoon that their primary AI infrastructure vendor was filing for Chapter 7 bankruptcy. The platform hosted their patient risk stratifica

AI Observability Beyond Logging: Trace Replay, Incident Forensics, and Cost Attribution
AI Observability Beyond Logging: Trace Replay, Incident Forensics, and Cost Attribution
22 Jun, 2026 | 11 Mins read

Traditional application observability focuses on three signals: request latency, error rates, and resource utilization. If the request returns a 200 in under two hundred milliseconds, the system is he

Open-source sustainability: who pays for the code everyone uses?
Open-source sustainability: who pays for the code everyone uses?
22 Jun, 2026 | 05 Mins read

A critical open-source library used by thousands of companies, including several Fortune 500 firms, is maintained by one person in their spare time. This is not a hypothetical. It is a description of

Tool Governance for MCP: Scoping Permissions Before They Drift
Tool Governance for MCP: Scoping Permissions Before They Drift
21 Jun, 2026 | 10 Mins read

When an AI agent can call external tools, the security boundary shifts from the model to the tool layer. The model generates a request to call a tool. The tool executes against real systems — reading