Overview
Lead the design and delivery of multi-agent systems on the Agora Orchestrator platform at MRI Software. Architect agent orchestration, author MCP servers, design RAG systems, and set the engineering standards the team follows.
What you'll do
- Design and implement multi-agent orchestration using LangGraph, including state, checkpointing, and human-in-the-loop.
- Author MCP servers, including tool schemas, authentication, error contracts, and versioning.
- Architect RAG systems, including retrieval strategy, reranking, hybrid search, and evaluation.
- Build and maintain the agent evaluation and observability stack.
- Deploy and operate agents on Azure Foundry Agent Service and Azure Functions.
- Establish agentic patterns and reference implementations for the team.
- Mentor junior engineers.
- Lead technical design reviews.
What you'll need
- Expert-level Python programming, including async, typing, packaging, and performance.
- Experience with LangGraph in production, including multi-agent orchestration, state management, checkpointing, and human-in-the-loop.
- Experience with LangChain, including tools, retrievers, and callbacks.
- Experience with FastAPI, including async, authentication, and streaming.
- Experience with Azure Foundry Agent Service, including provisioning, threads, and tool wiring.
- Experience with Azure Functions.
- Experience with AI agentic orchestration and agentic patterns, including planner/executor, supervisor, reflection, tool use, and multi-agent collaboration.
- Experience with MCP server design and authoring, including tool schemas, authentication, error contracts, and versioning.
- Experience with RAG architecture, including retrieval strategy, hybrid search, reranking, and evaluation.
- Experience with prompt engineering using structured evaluation frameworks such as LangSmith, Promptfoo, or equivalent.
- Experience with agent observability, including tracing, App Insights, token/cost telemetry, and failure-mode analysis.
- Experience with PyTorch, Hugging Face Transformers, sentence-transformers, scikit-learn, NumPy, and Pandas.
- Experience with identity technologies including Okta, OAuth2, and JWT, including client_id propagation and scopes.
- Experience with system design for multi-tenant SaaS, idempotency, retries, and queueing.
- 5+ years building software in Python.
- 2+ years building production LLM or agentic systems.
- A track record of leading technical design and mentoring engineers.
Details
- Location: Baroda, India.
- Hybrid working arrangements with 3 days working from office in a week.
Read the full description and apply on the company’s own careers page.