Agent Architect – Agentic Systems
On-siteMohali, Punjab, India
Job Summary
Define blueprints for multi-agent workflows, including reasoning loops, tool integration, and memory systems. Architect scalable agentic systems using LangGraph, MCP, and A2A messaging while managing context pipelines via RAG and knowledge graphs. Build frameworks connecting LLMs, APIs, and observability layers, then implement guardrails for safety, compliance, and brand alignment. Mentor AI engineers on trade-offs between accuracy, cost, and latency. Partner with product teams to align agent design with enterprise outcomes. Design production-ready architectures for Fortune 500 clients in regulated industries, ensuring systems are trustworthy, explainable, and scalable.
Required Qualifications
- 6–10 years in AI/ML engineering, systems architecture, or enterprise software design
- Deep knowledge of LLM architectures, orchestration frameworks (LangChain, LangGraph, LlamaIndex), and agent design patterns
- Strong understanding of context engineering, RAG pipelines, and vector/knowledge databases
- Experience with multi-agent orchestration frameworks (MCP, A2A messaging, AgentBridge)
- Proficiency in Python and AI development stacks
- Familiarity with cloud-native architecture (AWS, GCP, Azure) and container orchestration (Kubernetes, Docker)
- Solid understanding of Responsible AI frameworks and compliance-driven system design
Desired Qualifications
- Background in Reinforcement Learning (RLHF, RLAIF, reward modeling)
- Experience in enterprise-scale deployments in BFSI, GRC, SOC, or FinOps
- Prior experience as a Solutions Architect, AI Architect, or Technical Lead
- Contributions to open-source frameworks in multi-agent systems or LLM orchestration
Hiring someone like this?
Get your role in front of qualified candidates on Sorce.