AI Solution Architect

Role Summary

This individual designs and validates cost-effective AI solutions that meet client and business requirements, translating complex needs into implementable architectures. Works primarily with development and delivery teams to ensure the design is built and delivered successfully, and supports the sales and offering teams as a technical advisor rather than a commercial owner. Hands-on work is situational — the ability to produce FDE-style mock-ups and prototypes is highly desired. 

• Design end-to-end AI solution architectures — data pipelines, model serving, integration, and security controls — against functional and non-functional requirements.  
• Define deployment strategy and the path from pilot to production; embed responsible-AI, privacy, and regulatory requirements in the design.  
• Work day-to-day with the development team through build — design reviews, backlog shaping, and unblocking technical decisions.  

• Build FDE-style mock-ups, prototypes, and thin vertical slices to de-risk a design or make a concept tangible for the client.  
• Prove out integration patterns, model behavior, and performance assumptions in code before the team commits to them.  
• Contribute to the codebase where the situation calls for it — spikes, reference implementations, accelerators — without becoming a permanent delivery resource.  

• Provide and review technical content for RFP/RFI responses, proposals, and SOWs; support pricing with defensible effort and run-cost inputs.  
• Design and run demos, proofs of concept, and pilots with explicit success criteria and a defined route to production.  
• Convert repeatable client patterns into productized offerings, accelerators, and reference architectures.  

• Languages & frameworks: Python (primary), SQL, JavaScript/TypeScript; FastAPI, LangChain or LlamaIndex, one agent framework.  
• AI/ML: LLM prompting and evaluation, RAG, embeddings and vector stores, fine-tuning, guardrails, inference cost and latency optimization.  
• Cloud & platforms: AWS, Azure, or Google Cloud AI/ML stack; containers (Docker, Kubernetes), IaC (Terraform), CI/CD.  
• Data: Data modeling, pipelines and ETL/ELT, streaming and batch patterns, lakehouse and warehouse platforms (Databricks, Snowflake, or equivalent).  
• Enterprise integration: REST/GraphQL APIs, event-driven and messaging patterns, SSO and identity, integration with ERP, CRM, and ITSM platforms.  
• Architecture & security: Reference and solution architecture documentation, non-functional design (scalability, availability, observability), data privacy and responsible-AI controls.  

• Bachelor's degree in computer science, engineering, or a related field, or equivalent practical experience.  
• 2–5 years of combined solution architecture and software development experience, including at least one design taken through to delivery.  

• Solid understanding of enterprise business applications — ERP, CRM, ITSM, data platforms — and the integration patterns that connect them.  
• Practical experience delivering AI/ML or generative-AI solutions into production, with working knowledge of a major cloud platform (AWS, Azure, or Google Cloud).  
• Command of modern AI patterns: RAG, agentic workflows, fine-tuning, evaluation, and guardrails.  
• Current hands-on ability to produce a credible prototype independently (Python plus a modern AI/agent framework).  
• Strong communication — able to explain a design to engineers, client stakeholders, and sales; willing to travel to client sites as needed.  
• Cloud or AI professional-level certifications (AWS/Azure/GCP architect or ML specialty).  
• Consulting, systems integrator, or ISV background supporting sales and offering teams from a delivery-side role.  
• Industry or enterprise functional depth (financial services, healthcare, manufacturing; finance, supply chain, service).  
• MLOps/LLMOps tooling experience, or forward-deployed and embedded engineering work inside a client's environment.  

• Architectural judgment — chooses the simplest design that meets the requirement and defends the trade-offs.  
• Business context — fits the design to the client's processes and application landscape.  
• Ambiguity tolerance — produces a workable design from incomplete inputs and refines it as facts arrive.  
• Influence without authority — aligns development, delivery, sales, and partners on one solution.  
• Builder's instinct — reaches for a prototype when a document would not settle the question.  

• Architected solutions reaching production on scope and on schedule.  
• Pilot-to-production conversion rate.  
• Reuse of reference architectures and accelerators across accounts.  
• Development-team and sales feedback on support quality.