16.07.2026 aktualisiert


verifiziert
Premiumkunde
100 % verfügbarAI Solution Architect & Forward Deployed Engineer | LLM · RAG · MCP · Cloud-Architektur
Berlin, Deutschland
Nur Remote
Über mich
Ich bin Jack Schmidt – AI Solution Architect & Forward Deployed Engineer mit 10+ Jahren Erfahrung. Ich baue produktionsreife AI-Systeme (LLM, RAG, KI-Agenten, MCP) und Cloud-Datenplattformen – von der Architektur bis zum Betrieb – und modernisiere komplexe Web-Plattformen mit Nuxt und Next.js.
Skills
AI AgentsRAGOpenAILangChainSupabaseKünstliche IntelligenzAmazon Web ServicesAmazon S3ArchitekturAutomatisierungTest AutomationSaasCloud ComputingDatenbankenContinuous Integration
Welcome to my profile
I'm an AI Solution Architect and Forward Deployed Engineer. I help organizations translate complex technical challenges into production-ready AI systems and data platforms, from discovery and architecture decisions through implementation and stable operation.
My foundation is 10+ years of fullstack product engineering. Today my work goes further: I embed generative AI directly into products and into the way teams work, from Retrieval-Augmented Generation (RAG) assistants and LLM-powered semantic search to autonomous AI agents and MCP servers.
Typical challenges I solve as an AI Solution Architect
- AI enablement: LLM integration, RAG systems, semantic search, vector databases (pgvector), AI agents, Model Context Protocol (MCP)
- Data platform design: ETL pipelines (Dagster), data warehouses, cloud-native architecture (AWS, Cloudflare, Supabase)
- Large-scale migrations (Vue 2 → 3, Nuxt 2 → 3, legacy stacks → modern TypeScript & serverless)
- Multi-tenant SaaS and enterprise integrations with security & compliance in mind (GDPR, ISO 27001, SOC 2)
- Headless commerce transformations (Shopify, MACH architecture)
My approach
- Architecture first — mapping the current landscape and designing scalable, future-proof patterns
- AI-native delivery — LLMs, vector databases, prompt engineering and evaluation suites for features that learn and adapt
- Iterative delivery — short, value-focused sprints with CI/CD and automated testing (Playwright, Vitest, GitHub Actions)
- Enable & empower — coaching teams on AI-native workflows so capability remains long after I leave
As a Forward Deployed Engineer I work at the interface between business and engineering, embedded with the client, deep in the product, with clear ownership of outcome and operation.
Open source & writing I maintain anchored (github.com/chafoo/anchored), an open-source framework for specification-driven verification of AI-generated code, and write about AI-assisted software engineering — e.g. "The Missing Half of Trust in AI Coding" on dev.to. Verifying and trusting AI output is, in my view, the core engineering challenge of the coming years.
Technology stack
I pick the stack to fit the problem, not the other way around. My base is the TypeScript ecosystem, extended by whatever solves the task best.
- AI & Data: OpenAI, LLM, RAG, AI agents, MCP, LangChain, Vercel AI SDK, embeddings, semantic search, pgvector, Supabase, PostgreSQL, ETL, data warehousing
- Frontend & Delivery: TypeScript, Nuxt/Vue, Next.js/React, Astro, TanStack, Node.js, Deno, Hono, Express
- Cloud & DevOps: AWS (Lambda, S3, Bedrock), Cloudflare, Vercel, Netlify, Docker, CI/CD
- Commerce: Shopify (Headless), Contentstack, MACH architecture
What I value
I love partnering with organizations that have ambitious ideas and want to move fast on high quality. Multilingual, culturally diverse environments are my home turf.
Do you have an AI challenge or a migration in mind?
Sprachen
DeutschMutterspracheEnglischverhandlungssicher
Projekthistorie
Embedded with Philips Sound to design and build a global data platform for aggregating, processing and analyzing international social media data — from architecture decisions through production operation. Owned the end-to-end solution architecture: robust ETL pipelines built with Dagster, a PostgreSQL data warehouse structured for reporting and analytics, and cloud infrastructure on Cloudflare. Developed a central MCP server with a lightweight Hono backend, giving international teams secure, unified and AI-ready access to company data.
Achievements
Architected and delivered a production-ready global data platform as a single source of truth for international teams. Built scalable ETL pipelines (Dagster) and a PostgreSQL data warehouse for analytics and reporting. Developed a central MCP server on Cloudflare enabling secure, controlled and AI-ready data access. Established the foundation for AI-powered analytics and reporting workflows across the organization.
Skills
Dagster, PostgreSQL, Cloudflare, Hono, MCP, TypeScript, ETL, Data Warehousing, OpenAI, Data Engineering, Cloud Architecture
Built comie.ai from scratch as a solo venture: a multi-tenant SaaS platform providing embeddable AI widgets for e-commerce websites — combining my two core fields, e-commerce and applied AI. Designed and implemented the complete architecture independently: Next.js frontend with the Vercel AI SDK, Supabase as central backend with PostgreSQL and pgvector as vector database for semantic retrieval (RAG), OpenAI for context-aware AI interactions, and a fully functional subscription and billing system via Stripe. Deployed on Vercel.
Achievements
Designed and shipped a production-ready multi-tenant SaaS platform end-to-end without a team. Built embeddable AI widgets consumable by third-party e-commerce sites. Implemented RAG-based, context-aware responses via OpenAI and pgvector similarity search. Implemented a complete subscription and billing flow with Stripe.
Skills
Next.js, TypeScript, Supabase, PostgreSQL, pgvector, RAG, Vercel AI SDK, OpenAI, Stripe, Multi-Tenant SaaS, Vercel
Solo engagement with end-to-end technical ownership of the international e-commerce platform redesign for Philips Sound. Acted as the technical bridge between the Philips brand team, the agency and external development partners. Owned concept, UI/UX and the full high-fidelity design, and built a scalable design system in Figma as the architectural foundation for implementation — including project steering from structure through developer handover, without a dedicated design team.
Achievements
Carried technical and conceptual ownership of an international commerce platform redesign as a solo engagement. Built a scalable design system and component library in Figma as the basis for implementation. Aligned three stakeholder groups (brand team, agency, external developers) on one clear technical direction.
Skills
Figma, Design Systems, System Design, Architecture Governance, Stakeholder Management, E-Commerce