ZAID FAROUQI

AI / LLM Engineer and Application Lead with experience building production-grade AI systems including RAG pipelines, LLM orchestration, MCP tool execution, and evaluation frameworks.

Designed and implemented end-to-end AI platforms with vector search, hybrid retrieval, real-time streaming, and multi-model LLM routing. Experienced in building agentic workflows, evaluation systems, and AI analytics for cost, latency, and reliability.

Strong software engineering background with a focus on scalable architectures, observability, and production-ready AI systems.

• Designed and built a production AI platform with RAG, evaluation, MCP tools, and analytics
• Implemented end-to-end RAG pipeline:
• query rewriting & expansion
• hybrid retrieval (dense + BM25)
• LLM reranking
• streaming responses with chunk attribution
• Built LLM orchestration layer:
• multi-provider model switching (runtime)
• token streaming & latency tracking
• cost estimation and usage analytics
• Developed MCP tool execution system:
• natural language → tool selection → execution
• argument extraction, validation, normalization
• backend API integration
• Implemented evaluation pipelines:
• retrieval metrics (MRR, nDCG)
• answer quality (LLM-as-judge)
• evaluation history tracking
• Built LangGraph agent workflows:
• graph-based execution
• multi-step AI flows
• Designed LLM analytics system:
• token usage, latency (avg/p95), error rates, cost
• Implemented AI streaming architecture (SSE)
• Implemented hybrid retrieval using dense embeddings + BM25 keyword search

• Leading for Impact: Leadership Boot Camp – Dale Carnegie
  Oct 2025