Specialized AI Agent for Microsoft Cloud Solution Architect (Azure Security) Interview Prep

I am seeking an expert AI Engineer to develop a high-fidelity Interview Preparation Agent. The goal is to prepare for a Cloud Solution Architect - Azure Security role at Microsoft. The agent must simulate a multi-stage interview loop, evaluate technical/architectural accuracy against Microsoft-specific frameworks (CAF, WAF), and provide actionable coaching. Detailed Scope of Work & Features: Multi-Stage Interview Simulation (Loops): Phase 1: HR/Behavioral: Focus on Microsoft Leadership Principles and Growth Mindset. Phase 2: Technical Deep-Dive: Interactive grilling on Azure IaaS, Security (WAF, Key Vault, Azure AD), Networking, and Kubernetes. Phase 3: System Design & Strategy: Architecture-level discussions focused on "Support for Mission Critical" scenarios. Context-Aware Evaluation (RAG): The agent must use Retrieval-Augmented Generation (RAG) to pull from specific Microsoft documentation (Azure Architecture Center, CAF, and WAF) to verify if my answers align with official Microsoft best practices. Voice-to-Voice Interface: Integration with OpenAI Whisper (STT) and ElevenLabs or OpenAI Voice (TTS) to allow real-time spoken mock interviews. Active Coaching & Scoring: The agent must provide a "Scorecard" after each session, grading me on: Technical Accuracy, STAR Method Structure, and Cultural Fit. Long-Term Memory: The agent must store a history of my mistakes and focus future sessions on weak areas (using a Vector DB like Pinecone or ChromaDB). Specific Deliverables (Must provide to complete project): Functional Web/Local App: A clean UI (Streamlit, Chainlit, or React) where I can upload my CV and the specific Job Description. Configurable Knowledge Base: A module where I can upload PDFs/URLs (Whitepapers, Case Studies) that the agent will use as its "Source of Truth." The "Prompt Engineering" Library: A documented set of System Prompts used for each interview persona (The "Hard" Technical Lead vs. the "Empathetic" Manager). Session Logging & Analytics: A dashboard or log file that exports my performance trends over time. Deployment & Documentation: A README file with instructions on how to run the agent locally (Dockerized or Python Environment) and how to update the LLM keys. Technical Stack Preferred: Orchestration: LangChain, CrewAI, or Microsoft Semantic Kernel. LLM: GPT-4o or Claude 3.5 Sonnet. Database: Vector DB for RAG and SQL/NoSQL for session memory. Voice: Integration with Real-time Speech APIs. Application Instructions (Filtering Bot Responses): Applications that do not include the following will be ignored: The Technical Question: "How will you handle 'Hallucinations' when the agent evaluates a complex Azure Networking scenario that isn't explicitly in the RAG documentation?" Portfolio: Links to previous AI Agent or RAG projects you have built. Tech Choice: Which orchestration framework (LangChain vs. Semantic Kernel) do you recommend for this specific Microsoft-aligned project and why?

Project ID: 40275477