We’re looking for an AI Solution Architect (freelance) to lead the design and delivery of enterprise AI systems — from AI-native applications and agentic workflows to the data foundations that make them reliable. You’ll own architecture end-to-end: shaping solutions with clients and stakeholders, setting standards, and staying hands-on enough to prove out the hard parts yourself. You’ll also help drive AI adoption across our own engineering practice, so teams ship faster and better.
This role suits someone who is equally comfortable in a whiteboard session with a client’s CTO and in a terminal debugging a RAG pipeline. What you’ll do * Design end-to-end AI architectures that map to real business goals, not just technical possibility. * Architect AI solutions across common enterprise patterns: copilots, RAG and document intelligence, conversational AI, workflow automation, and multi-agent systems. * Design AI agents with tool calling, memory, orchestration, and human-in-the-loop controls — and know when not to make something autonomous. * Build AI-ready data layers: vector stores, semantic search, knowledge graphs, metadata, and RAG pipelines grounded in enterprise knowledge. * Integrate AI with systems of record (ERP, CRM, HRIS, ITSM, finance, collaboration tools) via APIs and event streams. * Drive AI adoption across the SDLC — AI-assisted design, coding, testing, review, documentation, and CI/CD — and measure whether it’s actually helping. * Provide technical leadership: architecture standards, governance, mentoring, and client engagement.
What you’ll bring * 8+ years in solution or software architecture, including 3+ years delivering production AI solutions (not just prototypes). * Hands-on experience with at least one major LLM platform: OpenAI, Anthropic Claude, Google Gemini, AWS Bedrock, or Azure OpenAI. * Practical experience with at least one agentic framework — LangGraph, LangChain, LlamaIndex, Semantic Kernel, or OpenAI Agents SDK. * Strong Python, plus solid experience on at least one major cloud (AWS, Azure, or GCP). * Hands-on experience building RAG systems and working with vector databases. * Experience integrating systems with APIs, SQL/NoSQL, and enterprise integration patterns. * Excellent stakeholder communication — you can explain tradeoffs to executives and engineers alike.
Nice to have * TypeScript/JavaScript for full-stack or agent-tooling work. * Knowledge graphs, data lakes/lakehouses, and metadata/semantic modeling. * Daily use of AI developer tools (GitHub Copilot, Claude Code, Cursor) with a point of view on where they help. * LLMOps/evaluation experience — prompt versioning, offline/online eval, guardrails, observability. * Familiarity with AI governance, security, and compliance (data residency, PII handling, model risk). * Prior consulting or client-facing delivery experience.
What success looks like * Solutions reach production and deliver measurable business value — adoption, cost saved, or cycle time reduced — not just demos. * Reusable AI capabilities and data foundations that later projects build on instead of rebuilding. * Measurable acceleration of software delivery from AI adoption across teams. * AI integrated into enterprise systems with governance, security, and compliance intact.
