Applied AI operator turning messy, real‑world problems into clean specs, dataflows, and working AI systems.
I work at the intersection of user/product data, architecture, and agentic workflows.

• Report pipelines over unstructured text (logs, JSON, documents)
• Automated insight extraction from large text corpora
• Document understanding and summarization for decision‑makers

• Text data normalization with custom taxonomies
• Product and catalog data modeling for analytics and search
• Vector‑based similarity search using VectorDBs

• Multi‑agent orchestration and agent architectures
• Autonomous workflow design for business processes
• Retrieval‑augmented and grounded flows (RAG) over enterprise data
• Context routing and tool routing between agents
• Human‑in‑the‑loop review, guardrails and policy enforcement

• Context engineering and prompt orchestration
• Fast prototyping from BRD to spec to POC
• Report and knowledge‑base management
• Knowledge‑aware AI experiences
• Enterprise AI search and vector‑backed retrieval, context rotation for long‑running threads

• LLM cost and token‑usage monitoring across workflows
• Context‑window and context‑rotation optimization
• Quality–cost tradeoff modelling (latency, accuracy, spend)
• Offline and online evaluations and telemetry for AI features

Problem → Analysis → Spec → Implementation → Iteration

• Start from the business problem (enterprise data, workflows, pipelines, architecture)
• Map data sources, failure modes, and success metrics
• Design clear specs, dataflows, and agent flows that can be implemented and tested
• Use automation and AI tooling to ship quickly, then iterate until scalable and production ready

• OpenAI API, Anthropic API, Google Gemini, Vertex AI
• Retrieval‑augmented generation (RAG), VectorDB stacks
• Context routing, selection, and rotation strategies

• LangChain, LangGraph, Agno, Google Agent SDK, Anthropic Agent SDK
• Multi‑agent patterns, tool‑calling agents, grounded reasoning flows
• API‑driven workflows, data pipelines, system integrations

• Python (numpy, pandas, FastAPI)
• SQL, MongoDB, Azure, BigQuery
• ChromaDB and other vector stores

• n8n, Microsoft Power Automate, UiPath

• GitHub, GitHub Copilot, Claude Code
• Azure, Azure DevOps, Docker, deployment workflows
• Next.js frontends for AI products and internal tools

• Shipping applied AI features into production, not just prototypes
• Deepening work in multi‑agent systems, context engineering, and enterprise data patterns

• Email: vfaraji89@gmail.com
• LinkedIn: https://www.linkedin.com/in/vahid-faraji-jobehdar/