AI Solution Architect – Vertex AI, LLMs & Multi-Agent AI

• Design and govern a multi-agent AI architecture for vulnerability triage and automated remediation using Vertex AI, GPT, and Claude.

• Vertex AI

• GCP AI Services

• Prompt Engineering

• Large Language Models (LLMs), including:

• GPT

• Claude

• Gemini

• Lang Chain

• Semantic Kernel

• Vector databases, including:

• Pinecone

• FAISS

• Vertex AI Embeddings

• Python

• Machine Learning (ML) Pipelines

• Deep expertise in:

• Multi-agent collaboration

• LLM judge and scoring models

• Autonomous decision-making

• Fallback logic

• Design secure and explainable AI outputs.

• Design multi-agent architecture, including:

• Triage agents

• Remediation agents

• Judge agent

• Build LLM orchestration using Vertex AI, Lang Chain, and agent frameworks.

• Implement prompt engineering, prompt chaining, and tool usage delegation.

• Define Retrieval-Augmented Generation (RAG) pipelines using secure coding knowledge packs.

• Design agent memory using:

• Vector databases

• Knowledge graphs

• Establish model evaluation frameworks covering:

• Accuracy

• False positives

• Latency

• Define model versioning, evaluation harnesses, and governance frameworks.