Architecture review and documentation of a multi-agent AI codebase | Fixed scope

**There is nothing to build.** The platform already exists and runs in a private beta. I need an architect who reads the code, works out how it really behaves, and explains it clearly to two audiences: engineers, and people who will never read code. **About the platform** AgentS is a multi-agent platform. Every request follows the same main path: 1. Request: someone asks for something by chat or voice. Each request belongs to exactly one workspace. 2. Graph: an orchestrator turns the request into a dependency graph of steps (a DAG) from a versioned template. 3. Steps: each node runs a versioned operation from a catalog, such as research, analysis, writing or video. Run state and an append-only event log live in Postgres, and Inngest schedules the workers. 4. Human gate: some steps pause until an authorized person approves. The wait is durable: it survives restarts and is recorded. 5. Deliverable: the finished work (a brief, a report, a video) is stored per workspace, and row-level security keeps each customer's data apart. **Stack** One TypeScript monorepo: SvelteKit front end, Hono API on Bun, Postgres with pgvector and row-level security, Inngest for durable execution, LLMs through an AI gateway, Deepgram for voice, and an MCP-based tool layer. The stack is already chosen: you will read it, not change it. **How the work runs** This is a starting point. If you would run it differently, say how in your proposal. 1. Kickoff call: we sign the NDA, you get read-only access, and I walk you through the product and the repository layout. 2. Read and map: you trace the main path end to end in the code and draft the architecture map. 3. Mid-point check: we review the draft together to catch wrong assumptions before you start writing. 4. Write: you write the two versions of the document and the honest assessment. 5. Walkthrough: you present the architecture in a recorded call, and we close open questions. **Deliverables** - Architecture map: one diagram of components and data flow and one of the main path, each with its editable source file. - Document for engineers: every claim points to the files behind it, so a new engineer can check it against the code. - Document for non-technical readers: a client or investor understands what the platform does and why it can be trusted, without jargon. - Honest assessment: what is solid, what is risky and what you would simplify. Blunt is better than polite. - Recorded walkthrough: a call of about an hour that walks through the map. We keep the recording. **Out of scope** Writing features or changing code, refactoring or infrastructure work, and any access to production systems, credentials or customer data. **About you** - You have designed or reviewed production systems built on LLMs or agents. - You know durable workflow engines (Inngest, Temporal or similar) and multi-tenant Postgres. - You write and draw clearly for both technical and non-technical readers. - You work in English. Spanish is a plus, not a requirement. **Engagement** Remote, fixed scope, fixed price. NDA before any access. You get read-only access to the repository, the existing docs (concepts, architecture notes and runbooks), a live demo of the beta, and direct time with me for questions. Full scope with a visual walkthrough: https://playgrounds.digitalhubassist.ai/agentsquad-architecture-review.html **To apply, send:** 1. A system built on LLMs or agents that you designed or reviewed, and what your role was. 2. A document or diagram you made that explains a complex system to others. 3. Your fixed quote for this scope, how many weeks you need, and when you could start. Start your proposal with the word "gate" so I know you read this far. Proposals that skip the three items above won't get a reply.