Tian Pan
Software Engineer
The Cold Start Trap in AI Products
AI features need user data to work, but need to work to attract users. Here's how to escape the cold start trap without burning months on ML before your product earns the right to it.
The Confidence-Accuracy Inversion: Why LLMs Are Most Wrong Where They Sound Most Sure
Frontier LLMs exhibit their worst calibration in the domains where users trust them most. Here's how to measure the problem and build systems that handle overconfident wrong answers before they cause real damage.
The Copyright Exposure in AI-Generated Content: A Risk Framework for Engineering Teams
LLM outputs can reproduce verbatim training data, and the output liability can land with you — not the model provider. A practical engineering framework for measuring copyright exposure, implementing controls that actually work, and understanding the limits of provider indemnification.
Cultural Calibration for Global AI Products: Why Translation Is 10% of the Problem
LLMs are fluent in dozens of languages but calibrated to one culture. Here's what translation misses and how to engineer around it.
Database Connection Pools Are the Hidden Bottleneck in Your AI Pipeline
AI workloads break the assumptions behind standard connection pool sizing. Here's the math, the failure modes, and the patterns that actually work.
Deadline Propagation in Agent Chains: What Happens to Your p95 SLO at Hop Three
User-facing latency constraints silently disappear as requests traverse multi-step agent pipelines. Here's the structural problem behind that behavior, how major frameworks handle it (poorly), and the deadline-propagation patterns that fix it.
The Demo-to-Production Failure Pattern: Why AI Prototypes Collapse When Real Users Arrive
Demos run on cherry-picked inputs, warm caches, and patient evaluators. Production gets adversarial queries, distribution-shifted requests, and users who abandon in 8 seconds. Here's the pre-launch methodology that closes the gap.
The Deprecated API Trap: Why AI Coding Agents Break on Library Updates
LLMs produce fluent, confident code referencing APIs that no longer exist. Here's what causes it, how to measure it, and the layered defenses that actually work.
Distributed Tracing for Agent Pipelines: Why Your APM Tool Is Flying Blind
Standard APM tools break on multi-step agent pipelines. Here's what purpose-built observability for AI agents actually requires — and the three metrics that tell you an agent is degrading before users notice.
Document AI in Production: Why PDF Demos Lie and Production Pipelines Don't
The gap between a working PDF demo and a reliable production pipeline is vast. Here's what breaks, how to detect it, and how to architect for 10,000+ documents a day.
Document Extraction Is Your RAG System's Hidden Ceiling
PDF-to-text pipelines silently discard tables, scramble reading order, and destroy section hierarchy before your embedding model ever sees the data. Here's how to find and fix the real failure layer in your RAG system.
Earned Autonomy: How to Graduate AI Agents from Supervised to Independent Operation
A framework for gradually expanding AI agent operational scope based on measured performance history, with rollback triggers and oversight mechanisms that prevent premature autonomy.