GEN AI

Virtusa; Bengaluru; Vor Ort; INR 1.800.000 - 2.700.000; Vollzeit

Zusammenfassung

Virtusa in Bengaluru seeks an experienced AI Full Stack Engineer to build autonomous software systems, integrating agentic AI, asynchronous backends, and real-time frontends.

You will design multi-agent workflows, memory architectures, and scalable backends with Python or Node.js, and develop responsive React UIs with HITL approvals.

Qualifikationen

• Production-grade AI/agentic applications experience.
• Strong background in memory architectures and RAG.
• Experience with asynchronous backends and real-time frontends.

Aufgaben

• Agentic AI: Design self-correcting multi-agent workflows using LangGraph, CrewAI, or AutoGen.
• Memory & RAG: Implement short/long-term memory, hybrid RAG, vector and graph databases (Neo4j).
• Full Stack Development: Build scalable backends with Python (FastAPI, AsyncIO) or Node.js/TypeScript; real-time frontends with HITL via WebSockets/SSE.
• Data & State: Manage session state, memory, and traces with PostgreSQL, Redis, Pinecone, Weaviate, Qdrant, Pgvector.
• LLMOps & Guardrails: Observability with LangSmith, Phoenix, OpenTelemetry; use NeMo and Instructor for guardrails.
• DevOps: Containerize and deploy with Docker, Kubernetes, CI/CD, GitHub Actions; leverage Cursor and Copilot.

Kenntnisse

Full-stack development

Asynchronous runtimes

System design

Agent failure handling

Ausbildung

B.S./M.S. in Computer Science or AI

Tools

LangSmith

Phoenix

OpenTelemetry

NeMo

Instructor

Docker

Kubernetes

GitHub Actions

Cursor

Copilot

Jobbeschreibung

Job Requirements

AI Full Stack Engineer

Seeking an experienced AI Full Stack Engineer to build and scale production-grade autonomous software systems, integrating enterprise architecture, agentic AI, asynchronous backends, and real-time frontends.

Key Responsibilities

• Agentic AI: Design self-correcting multi-agent workflows (React, Plan-and-Execute, Reflexion), MCP-based tool orchestration, and pipelines using LangGraph, CrewAI, or AutoGen.
• Memory & RAG: Implement short/long-term memory, hybrid RAG, vector databases, and graph databases (Neo4j).
• Full Stack Development: Build scalable asynchronous backends using Python (FastAPI, AsyncIO) or Node.js/TypeScript; develop real-time React/Next.js frontends with execution feeds and human-in-the-loop (HITL) approvals using WebSockets/SSE.
• Data & State: Manage session state, memory, and traces using PostgreSQL, Redis, Pinecone, Weaviate, Qdrant, and Pgvector.
• LLMOps & Guardrails: Implement observability with LangSmith, Phoenix, and OpenTelemetry; use NeMo and Instructor to enforce guardrails and prevent tool hallucinations.
• DevOps: Containerize and deploy with Docker, Kubernetes, CI/CD, GitHub Actions, and AWS/GCP/Azure; leverage AI coding assistants such as Cursor and Copilot.

Key Technologies

AI/Agents: LangGraph, CrewAI, AutoGen, MCP, ReAct, OpenAI/Anthropic APIs, Ollama, vLLM.; Backend/Frontend: Python, FastAPI, AsyncIO, Node.js, TypeScript, React, Next.js, Tailwind, WebSockets, SSE.; Databases: PostgreSQL, Redis, Neo4j, Pinecone, Weaviate, Qdrant, Pgvector.; DevOps/LLMOps: LangSmith, Phoenix, OpenTelemetry, NeMo, Instructor, Docker, Kubernetes, AWS, GCP, Azure, GitHub Actions.

Qualifications

• Full-stack development experience, including 2–3 years building and deploying production-grade LLM/agentic applications.
• Strong command of asynchronous runtimes, system design, and agent failure modes, including loops, context loss, and hallucinations.
• B.S./M.S. in Computer Science, AI, or equivalent practical experience.

Work Experience

15-30Years