AI Systems Coach / Technical Advisor — 2 /Week Sessions (90 Days)

AI Systems Advisor / Technical Mentor — 2 Sessions Per Week (90 Days)Agents, n8n, Claude/OpenAI APIs, Production AI Systems, Productized AI Services IMPORTANT — Please read before applying We require a vendor who can issue a proper commercial invoice, operates as a registered business (LLC, Ltd, Inc, or equivalent), and can provide basic company documentation if requested. Individual freelancers operating without a registered business entity will not be considered regardless of skill level. This is a compliance requirement, not a preference. Objective Over the next 90 days, I want to compress several years of AI systems learning into an intensive apprenticeship. I learn by building. I am looking for someone who will review my systems, tell me where they are weak, assign projects between sessions, and pressure-test my work on the next call. Areas I want to develop: • Multi-agent architecture (orchestrator + worker patterns) • Agentic workflow design • n8n for workflow automation and agent orchestration • Frontier model APIs (Claude, OpenAI, Gemini) — tool use, structured outputs, memory management and model selection • Retrieval systems, vector databases, long-term memory, and MCP integrations • Evaluation frameworks — measuring whether AI systems actually perform in production • Observability and debugging (Langfuse, Braintrust, tracing, logging, failure analysis) • Context window management, prompt versioning, caching, and cost optimization • APIs, webhooks, authentication, and external tool integration • Deployment and production infrastructure (Docker, Railway, Render, Supabase, Postgres) • Decomposing real business operations into agent-executable workflows • Designing and selling productized AI services • Understanding real-world failure modes: hallucinations, broken handoffs, latency issues, context loss, and cost blowouts Who I'm looking for You or your firm has built and sold productized AI services. Clients have paid recurring revenue for AI-powered deliverables your company built and maintained. You operate at the systems level. Your firm has designed production AI systems and multi-agent pipelines. You've dealt with actual failures, hallucinations at scale, broken handoffs, cost overruns, and reliability problems—and you've architected around them. You teach through artifacts. Workflows. Repos. Prompt libraries. Loom walkthroughs. Code reviews. Real systems. I do not need slides or high-level frameworks. I want reps with someone who has done this commercially. You are direct. If my architecture is weak, say so. If my prompts are bad, say so. If my assumptions are wrong, challenge them. I'm not looking for encouragement. I'm looking for acceleration. Engagement Structure • Two calls per week (60–90 minutes) • Between-session build assignments • Async feedback via Loom, voice notes, or written comments • Review of code, prompts, workflows, and architecture decisions • 30 / 60 / 90-day milestones Expected outcome: By the end of the engagement, I expect to have designed, built, deployed, and documented at least one commercially viable AI system or productized AI service. Budget $5,000–$7,000 for the full 90-day engagement. Fixed-price with milestone structure preferred. To apply, answer these four questions only. No cover letter. What is your company's registered name and country of incorporation? (Required for vendor onboarding.) Describe one productized AI service your company has built and sold. What did it do? Who paid for it? What stack did you use? What broke in production? How did you fix it? What does your n8n or agent orchestration work actually look like? Share something I can evaluate. Workflow screenshots, GitHub repos, Loom walkthroughs, prompt libraries, or architecture diagrams are preferred. Generic portfolio links without specifics will not be reviewed. If we worked together for 90 days, what would you have me build during: • Days 1–30 • Days 31–60 • Days 61–90 Applications without answers to all four questions will not be reviewed.