AI Systems Consultant

About this role I am the founder of a bespoke software development consultancy. We build custom software for clients in construction, manufacturing, facilities management, and retail. We have a team of 15+ people. I have spent several months designing a complete internal operating system — an AI-powered platform that connects every part of my business into one place. I plan to build it over the next 14 weeks with my development team. Before the build starts, I want an experienced AI consultant to review the full specification, validate my thinking, identify any gaps or risks, and give me confidence that what my team builds will actually work the way I am imagining it. This is not a project where I need someone to do the design work. That work is already done. I need someone who can read what I have built, stress-test it, and either confirm it is solid or tell me what needs to change before the first line of code is written. What the system is The platform connects: AI-powered lead qualification and outreach, meeting intelligence, proposal generation, project delivery, client account management, finance and cashflow, staff performance monitoring, and a partner referral network. The full specification covers 25 screens, 41 database tables, 10 third-party integrations, and 4 AI prompts. The client management layer handles three commercial modes in one account view — new business, active projects, and managed service retainers. The AI layer uses Anthropic Claude for lead qualification, email drafting, reply handling, proposal generation, weekly operations reports, and yearly staff performance review packs. Every prompt is versioned and measured. The system learns from edit patterns and conversion data monthly. The technical stack is FastAPI + PostgreSQL + React + Celery/Redis. The integrations include Apollo, SendGrid, Gmail, Outlook, Calendly, HubSpot, Jira, SignNow, and Xero. What I will share with you • A complete master specification — 40+ pages covering every screen, every database table, every integration, every AI prompt, every role and access rule, and the full 14-week build order • Interactive HTML demos of every screen in the system — the full OS demo, the partner portal, the CRM, and the performance screens • A system flow diagram showing every step from lead acquisition through deal won, delivery, client CRM, and the intelligence learning loop • The engineering rules, the complete environment variable reference, and the Cursor AI build guide the dev team will use What I need you to do A structured review across five areas: 1. AI architecture review — are the four prompts correctly designed? Is the qualification scoring model sound? Is the learning loop structured in a way that will actually improve outputs over time? Are there better approaches I should consider? 2. Data model validation — do the 41 database tables make sense? Are there missing tables or columns that will cause problems? Are the relationships correct? Is the multi-tenant architecture sound? 3. Integration design review — are the 10 integrations sequenced correctly? Are there failure modes I have not accounted for? Is the stub-first approach right? Are there better patterns for this stack? 4. Scalability assessment — will this architecture support 10× the current scale without a rebuild? Are there structural decisions being made now that will become expensive constraints in 18 months? 5. Risk and gap identification — what have I not thought of? What assumptions are baked in that may not hold in production? What does a system like this typically get wrong? What I need you to deliver • A written review document covering all five areas — not a slide deck, a proper written assessment with a clear verdict on each component • A prioritised change list — Critical (must fix before build) / Important (should fix) / Suggested (can address during build) — each with reasoning and a specific recommendation • A 20–30 minute video walkthrough of your key findings so I can share it with my dev team • A 60-minute live Q&A call with me and my two CTOs before the build starts • Written sign-off confirming the system is ready to build once any agreed changes are made Who I am looking for • Demonstrable experience designing or reviewing AI-powered business systems — not just prompt engineering, but full system architecture with AI as one component among many • Strong understanding of multi-tenant SaaS architecture • Familiarity with FastAPI, PostgreSQL, and Celery or an equivalent async task queue • Experience with the Anthropic Claude API or equivalent LLM integration in a production system • Someone who gives direct, honest feedback — I do not need validation, I need a rigorous review • Strong written English — the main deliverable is a written document