Page 111
12 articlesVision Inputs in Production AI Pipelines: The Preprocessing Decisions Nobody Documents
How image resolution, compression artifacts, OCR preprocessing, and aspect-ratio handling silently degrade vision model accuracy in production — and the normalization pipeline that separates model failures from input failures.
The Warranty Problem: Who Pays When Your AI Feature Is Wrong?
Software warranties assumed deterministic behavior — AI features break that assumption. A practical guide to the liability, insurance, and contract gaps engineering teams face when shipping non-deterministic systems.
When Your Agents Disagree: Consensus and Arbitration in Multi-Agent Systems
How to resolve conflicting outputs from peer AI agents when there's no ground truth — covering majority voting, confidence weighting, judge models, and when to surface disagreement to users rather than hide it.
Write-Ahead Logging for AI Agents: Borrowing Database Recovery Patterns for Crash-Safe Execution
Database WAL patterns map directly to AI agent workflows — an execution journal that logs intent before action and outcome before advancing enables skip-replay recovery, exactly-once side effects, and deterministic resumption after mid-workflow crashes.
Capability Probing: How to Map Your Model's Limitations Before Users Do
Map your LLM's failure boundaries before deployment using probe suites, capability matrices, canary prompts, and a probe-to-regression pipeline that catches silent regressions across model upgrades.
CZ's 'Freedom of Money': from a Jiangsu Boy to a Crypto Empire - Chapter-by-Chapter Summary
A comprehensive chapter-by-chapter breakdown of Changpeng Zhao's autobiography, 'Freedom of Money.' From a rural Jiangsu village to a Canadian immigrant, from a Wall Street coder to founding the world's largest crypto exchange, and his journey through a guilty plea, prison, and newfound freedom—all 25 chapters detailed.
The Hidden Token Tax: How Overhead Silently Drains Your LLM Context Window
System prompts, tool schemas, and chat history silently consume 30-60% of your LLM context window before user content arrives — here's how to audit and cut the overhead.
The Infinity Machine: How Demis Hassabis Built DeepMind and Chased AGI
From chess prodigy to Nobel Prize co-winner, Demis Hassabis built DeepMind into the world's most ambitious AI research lab. Sebastian Mallaby's biography traces the scientific breakthroughs, corporate battles, and existential dilemmas behind the quest for artificial general intelligence.
The Accuracy Threshold Problem: When Your AI Feature Is Too Good to Ignore and Too Bad to Trust
Deploying an AI feature at 70–85% accuracy creates a uniquely dangerous zone: good enough to attract habitual use, bad enough to cause visible failures that collapse user trust. Here's what the research says about why this zone is so treacherous and how to design your way out of it.
Adversarial Agent Monitoring: Building Oversight That Can't Be Gamed
Single-layer LLM-as-judge monitoring fails over 52% of the time against sophisticated agents. The four-layer defense stack — behavioral fingerprinting, action auditing, multi-monitor consensus, and tool-layer constraints — that holds up in production.
Why Agent Cost Forecasting Is Broken — And What to Do Instead
Traditional cost forecasting fails for AI agents because execution paths are stochastic, not deterministic. Learn decision-loop cost modeling, Monte Carlo simulation, and the guardrail patterns that make agent spend predictable.
Agent-Friendly APIs: What Backend Engineers Get Wrong When AI Becomes the Client
Most REST APIs silently break when AI agents become the client — ambiguous errors cause retry loops, offset pagination corrupts traversals, and request-count rate limits collapse under multi-agent coordination. Here's what to fix and why it matters.
About Tian Pan
I'm Tian Pan, an engineer-founder focused on agentic engineering — building autonomous AI systems and scaling engineering teams. I write practical guides on system design, technical leadership, and shipping with AI agents. Previously an early engineer at Uber, Brex, and IoTeX.
