Nuaav
Website:
nuaav.com
Job details:
Job Description – Senior Agentic AI Engineer (AI workflows on Google Cloud Platform (GCP)
Location: Remote / India || Experience: 4-8 years
Quick Technical Check
- Do you have hands-on experience building production workflows or advanced prototypes natively using the Anthropic/Claude API or Claude Code?
- Do you have hands-on experience building, publishing, or integrating Model Context Protocol (MCP) servers?
- You are proficienct with Python backend engineering (specifically writing asynchronous code, building production-ready REST APIs with FastAPI, and managing ORMs like SQLAlchemy)?
- You are hands-on experience implementing multi-agent orchestration patterns (such as supervisor/worker, planning, reflection, or self-correction workflows) in a production environment?
- You have production experience working with PostgreSQL (or similar SQL databases) alongside modern vector databases (e.g., Pinecone, Weaviate, Qdrant, or pgvector)?
- You have hands-on experience deploying applications, services, or AI workflows on Google Cloud Platform (GCP) (e.g., using Cloud Run, Vertex AI, GKE, or Pub/Sub)?
Role details -Relevant Experience: 4-7 years
About the Role
We're building production multi-agent systems on top of Claude, and we need an Agentic AI Engineer who lives inside the Claude ecosystem and can ship backend APIs. You'll design and build Claude-native agents, skills, subagents, hooks, and rules — and also build the FastAPI endpoints and MCP servers that those agents call as tools. You'll use Claude Code daily and deploy on our Python + GCP stack.
What You'll Build With (Claude Stack)
- Claude Code: as the primary development environment, including subagents, slash commands, and project setup
- Claude Agents & Subagents: designing role-specialized agents with scoped tools and context
- Claude Skills: authoring reusable SKILL.md packages that extend agent capabilities
- Hooks: pre/post tool-use hooks, prompt hooks, and workflow automation triggers
- Rules & Project Memory: CLAUDE.md, project rules, and persistent context engineering
- MCP (Model Context Protocol): building and integrating MCP servers as agent tools
- Anthropic API: function/tool calling, structured outputs, streaming, prompt caching, extended thinking
- Multi-agent orchestration: supervisor/worker, planner/executor, reflection, and handoff patterns
Responsibilities
- Design and ship Claude-native agents and subagents with clear roles, tools, and guardrails
- Author Claude Skills (SKILL.md) that package reusable capabilities for agents
- Build hooks for tool-use validation, safety checks, logging, and workflow automation
- Define project rules, CLAUDE.md, and memory strategies for consistent agent behavior
- Build backend APIs in Python/FastAPI that serve as tools for agents — clean contracts, idempotency, structured errors, streaming where needed
- Build and integrate MCP servers that expose internal APIs, databases, and tools to agents
- Implement multi-agent orchestration patterns: planning, delegation, reflection, self-correction
- Develop memory systems, evals, and feedback loops for agent improvement
- Implement guardrails, monitoring, and safety mechanisms for production agent behavior
- Use Claude Code daily for both building agents and as a development tool
- Collaborate with the Backend and GCP Application Engineers to deploy agents on Python + GCP
- Evaluate agent performance through benchmarks, traces, and real-world testing
Requirements
- 3–5 years in AI/ML engineering, with at least 1–2 years focused on LLM applications
- Hands-on experience with the Claude ecosystem: Claude Code, agents, subagents, skills, hooks, rules, and MCP
- Production experience with the Anthropic API: tool use, structured outputs, streaming, prompt caching
- Experience building multi-agent systems (with Claude-native patterns and/or LangGraph, CrewAI, AutoGen)
- Strong Python skills, including backend API development with FastAPI (or Starlette / Flask) — async patterns, REST design, authentication, error handling
- Comfortable designing API contracts that double as agent tools
- Familiarity with PostgreSQL (or similar SQL DB) and ORMs (SQLAlchemy or equivalent)
- Solid grasp of prompt engineering, context engineering, and tool design for agents
- Familiarity with vector databases (Pinecone, Weaviate, Qdrant, pgvector)
- Understanding of agent design patterns: ReAct, Plan-and-Execute, supervisor/worker, reflection
- API integration and workflow automation experience
Nice to Have
- Built and published MCP servers (open-source or internal)
- Authored public Claude Skills or contributed to the Claude/Anthropic ecosystem
- Experience with Docker and containerized deployments
- Background task / queue experience (Celery, Arq, Redis, Pub/Sub)
- RAG pipelines and retrieval system design
- Agent evaluation frameworks (LangSmith, Braintrust, custom eval harnesses)
- Open-source contributions to agent projects
- Experience deploying agents on GCP (Cloud Run, Vertex AI Agent Builder)
Nuaav - Company Overview
Nuaav is a boutique consulting firm that positions itself as a strategic partner for corporate leaders, focusing on delivering real business results and driving impactful change in complex and rapidly evolving business environments. Unlike large-scale consulting firms, Nuaav brings the boutique advantage—specialized expertise, personalized attention, agility, and tailored solutions—ensuring higher service quality and true alignment with client needs.
Our core service areas include:
- Product Engineering – "Crafting Innovative Solutions": End-to-end product lifecycle development with emphasis on innovation, reliability, scalable DevOps practices, and integration of emerging technologies such as AI, IoT, and Blockchain.
- Data Systems – "Powering Insights with Precision": Enterprise-scale data architecture, integration, governance, real-time processing, and big data solutions to fuel smarter decision-making.
- AI and Machine Learning – "Intelligence for the New Age": Predictive analytics, automation, and intelligence solutions to optimize strategy, performance, and customer engagement.
- User Experience (UX) – "Design with Purpose": Intuitive, research-driven design solutions to enhance customer experiences and retention.
At Nuaav, we combine deep technology expertise with business acumen to create integrated solutions that support organizations from initial product conception through to scalable data-driven growth.
Share details on artee.puri@nuaav.com
Click on Apply to know more.