I'm a Software Engineer building production-grade Generative AI systems and scalable backend architectures.
My work focuses on taking LLM-based ideas from prototype to reliable, observable, and cost-efficient systems.

I enjoy working where GenAI meets backend engineering — agents, RAG pipelines, APIs, and infrastructure that actually scales.

• LLM-powered applications using LangChain-style abstractions
• Agentic systems built with LangGraph (stateful workflows, tool orchestration)
• Retrieval-Augmented Generation (RAG) pipelines
• MCP-based servers for structured model–tool communication
• Designing scalable, fault-tolerant backend services
• Performance, cost, and latency optimization for AI workloads

• Designing multi-step agent workflows (planner → executor → tools)
• Building RAG systems with embedding pipelines, vector stores, and reranking
• MCP server patterns for safe & modular tool exposure
• Caching, batching, and async execution for LLM APIs
• Observability, retries, and graceful degradation in AI systems

Languages: Python, JavaScript, TypeScript
GenAI Frameworks: LangChain, LangGraph
Protocols & Tooling: MCP Servers, Tool Calling
Backend: FastAPI, Flask, Node.js, Express
Datastores: MongoDB, MySQL, Redis
Vector Search: FAISS / Pinecone-style systems
Infra: Docker, AWS, GCP

• 🔹 RAG-Based Knowledge Assistant
  Built a document-aware LLM system using LangChain pipelines, vector search, and reranking.

• 🔹 Agentic Workflow Platform (LangGraph)
  Implemented stateful AI agents with branching logic, tool execution, and memory persistence.

• 🔹 MCP-Enabled GenAI Backend
  Designed an MCP server exposing secure tools and APIs for LLM-driven workflows.

Focused on real-world reliability, not demos.

• LinkedIn: https://linkedin.com/in/arnavanand710
• Email: arnavanand710@gmail.com