AI Architect

You own the technical shape of systems other people must trust with real decisions. The model is the least stable part of your architecture; design for that. This playbook is anonymised from building document intelligence and knowledge systems in regulated and engineering contexts.

21stage document intelligence pipeline in production
5 daysgoverned knowledge hub, idea to live
150+documents indexed with cited answers
2public repos showing the patterns

The job in one sentence

Design AI systems whose answers can be checked, whose failures are visible, and whose models can be swapped without renegotiating trust.

Architecture principles

  1. Evidence first. Every answer carries its source. Citations and provenance are not features; they are the architecture's contract with its users.
  2. Abstain by default. When the corpus cannot answer, the system says so plainly. A confident wrong answer costs more trust than a hundred honest abstentions.
  3. Human gates are components, not process bolted on afterwards. Design the review queue, its interface and its capacity into the system the way you design the database.
  4. Pin to workflows, not models. Models change monthly; the user's workflow does not. The model sits behind an interface you can swap; the workflow contract is what you architect.
  5. Audit trail from day one. Who asked, what was retrieved, what was answered, who reviewed. In regulated contexts this is the difference between a pilot and a product.

The reference shape

Buy by default, build the differentiator

Buy or partner for anything the market does well. Build only where data sensitivity or genuine differentiation demands it, and know which of the two is your reason before writing code. The corollary is honest scoping: most organisations need governed access to good tools far sooner than they need a custom model. The executive version of this call is the three-question test in the CTO playbook.

The model is an implementation detail. The evidence chain is the product.

Evals are the contract

Security posture, stated honestly

The metrics that matter

Leading (predicts trust)Lagging (proves trust)
QualityEval pass rate per change, retrieval precision on the golden setCitation coverage sustained in production, incidents from wrong answers
OperationsCost per query, review-queue depth, gateway error ratesModel swaps completed without user-visible change, audit completeness

What not to claim

Questions to ask before you start

The patterns are public: AskLondon (cited answers, honest abstention) and Engineering the Box (the appliance architecture).

Designing an AI system that has to be trusted? Start with the advisory services or email hello@krishnachodipilli.com.