Essays · Operations Intelligence, AI and Quant
Hossein Narimani — Writing
In-depth writing on quant system design, operational AI, SaaS architecture, custom AI agents and founder execution systems.
OHLCV Feature Pipeline Architecture: Designing Feature Registries and Caches for Scalable Trading Systems
Most trading systems do not fail because of the model. They fail because every component interprets market data differently.When a feature has one value in a backtest and another in live trading, the problem is not intelligence. It is architecture.The Real...
Read more →
AI Agent Scaling Infrastructure Costs: Why Architecture Beats Model Choice
The Real Problem: Agent Costs Scale Exponentially, Not LinearlyMost teams pricing an AI agent look at one number: cost per token. That's the first...
Read more →
Are AI Agents Expensive to Implement? A System Architect's Honest Cost Breakdown
The real question isn't "are AI agents expensive?"Expensive compared to what? Compared to a human operator working around the clock? Compared to...
Read more →
AI-Driven Go-to-Market Strategy: The System Most Founders Get Wrong
The Real Problem: AI Gives You Speed, Not StrategyMost founders approach AI-driven go-to-market with a broken assumption. They believe the right...
Read more →
5 Common Mistakes in AI Agent Orchestration (And the Architecture That Fixes Them)
The Real Problem: Orchestration Doesn't Fail. The Design Does.When a multi-agent system breaks in production, the first explanation you hear is...
Read more →