# GenAI solution architecture and pre-sales

- Period: Oct 2024 – 2025
- Domain: Enterprise banking and telecom clients (Banking and financial services, Telecom)
- Role: Solution architect for proofs of concept and executive briefings
- Key result: 3, enterprise client wins, from 10+ proofs of concept
- Capabilities: Solution architecture and delivery; Knowledge, retrieval and context
- Stack: Azure OpenAI, Azure AKS, Azure AI Search, AWS Bedrock, AWS SageMaker, AWS Lambda, RAG
- URL: https://farhanbaluch.com/work/genai-solution-architecture

Azure and AWS reference architectures for RAG, agent hosting and guardrails, adapted per client into proofs of concept, plus executive briefings that turned interest into scoped work.

**AI angle.** Most enterprise AI work stalls between "this is interesting" and "this is approved". Reusable blueprints for RAG, agents and guardrails let each proof of concept start from a sound architecture, and the briefings turned leadership questions into pilots with a clear measure of success.

**What this shows:**
- I can design GenAI architectures on both Azure and AWS, including guardrails and hosting.
- I can explain agentic AI to executives and turn it into scoped, winnable work.

### Problem

Clients wanted to know what generative AI could do for them, but every proof of concept started from scratch, and leadership had no common picture of what an agent system needs in production.

### What I built

- Reference architectures on Azure and AWS for RAG, agent hosting and guardrails, which pre-sales adapted per client.
- RAG assistants and agent workflows as proofs of concept for enterprise banking clients.
- Executive briefings with live demos, ending in scoped proofs of concept with agreed success measures.

### Outcome

- The architectures backed 10+ proofs of concept and 3 enterprise client wins.