We are looking for a Senior AI Solutions Architect – Agentic AI & Multi-Agent Systems to architect and help build a production-grade AI Manager for our business.
This is not a chatbot, basic RAG implementation, or simple AI automation project.
Our internal software and AI team already has experience with LLMs, RAG, APIs, and standard automation. We are looking for a senior AI architect who has personally designed and built advanced agentic or multi-agent systems in real production environments.
You will work closely with our internal technical team to define the architecture, challenge existing assumptions, identify technical risks, and guide the development of a scalable, secure, reliable AI management system.
What We Are Building
We want to build an AI Manager that can operate like a digital business executive.
The AI Manager should be able to:
- Understand business goals, SOPs, policies, products, and historical decisions
- Access authorized live business data
- Analyze business situations
- Break complex objectives into tasks
- Plan and delegate tasks to specialized AI agents
- Coordinate multiple AI agents
- Review and validate agent outputs
- Use approved tools and APIs
- Make evidence-based decisions
- Detect uncertainty and conflicting information
- Request human approval for high-risk actions
- Execute approved actions autonomously
- Verify results after execution
- Maintain useful memory and decision history
- Maintain an audit trail of important decisions and actions
- Recover from failures and incomplete tasks
Potential specialized agents may include:
- Analyst Agent
- Research Agent
- SEO Agent
- Marketing Agent
- Product Agent
- Operations Agent
- Execution Agent
- Review / Critic Agent
The architecture is not fixed. We expect you to challenge this model and recommend a better approach where appropriate.
Key ResponsibilitiesAI Solution Architecture
- Design the overall architecture for a production-grade agentic AI platform.
- Determine when to use single-agent, multi-agent, deterministic workflow, or hybrid architectures.
- Define agent responsibilities, orchestration, communication, state, and execution flows.
- Design scalable LLM and tool-calling infrastructure.
- Define the interaction between business knowledge, memory, live data, business rules, and AI reasoning.
Multi-Agent Orchestration
- Design complex multi-step agent workflows.
- Build planning, delegation, coordination, review, and verification mechanisms.
- Define how specialized agents communicate and share context.
- Design critic/reviewer mechanisms to validate AI-generated outputs.
- Handle agent failures, conflicting outputs, uncertainty, and retries.
Tools, APIs & Business Systems
- Design secure tool-calling and API execution architecture.
- Integrate AI agents with authorized business systems and live data.
- Implement permission-based access and least-privilege controls.
- Ensure AI-generated actions are validated before execution.
- Design post-execution verification and rollback/failure mechanisms.
AI Security & Governance
- Design safeguards against prompt injection and malicious instructions.
- Prevent unauthorized AI access to business systems.
- Implement role-based permissions and action-level controls.
- Design risk-based human approval workflows.
- Establish audit logging for decisions, tool calls, approvals, and actions.
- Ensure sensitive systems and databases are not directly exposed to LLMs.
AI Evaluation & Reliability
- Design evaluation frameworks for AI agents and agent workflows.
- Measure task completion, decision quality, reliability, tool usage, and failure rates.
- Build automated and human evaluation mechanisms.
- Establish monitoring, tracing, logging, and observability.
- Design fallback and recovery strategies.
Technical Leadership
- Work closely with our existing AI and software engineering team.
- Review existing technical architecture and identify gaps.
- Create architecture diagrams, technical specifications, and implementation plans.
- Guide engineers during implementation.
- Conduct architecture and code reviews where required.
- Make key technical decisions around AI infrastructure and orchestration.
- Help transition systems from prototype to reliable production deployment.
Required Skills & Experience
Strong hands-on experience with:
- Agentic AI
- Multi-Agent Systems
- Generative AI / LLMs
- AI Solution Architecture
- AI Agent Orchestration
- LLM Tool Calling / Function Calling
- API & Business-System Integration
- AI Security
- AI Evaluation
- Human-in-the-Loop Systems
- AI Observability
- Memory / State Management
- Production AI Deployment
- Workflow Orchestration
Preferred Technologies
Experience with one or more of:
- LangGraph
- LangChain
- MCP
- OpenAI APIs
- Anthropic APIs
- Python
- FastAPI
- Vector Databases
- Event-driven architectures
- Docker / Kubernetes
- AWS / GCP / Azure
We care more about architecture, engineering judgment, and production experience than any particular framework.
Proven Production Experience Is Required
Candidates should have personally worked on advanced Agentic AI or Multi-Agent systems.
Your application should ideally include:
- A similar AI agent or multi-agent system you personally designed/built
- Company/client/project name, where disclosure is permitted
- What the system actually did
- Production vs POC/prototype status
- Your personal contribution
- Team size and your role in the project
- Architecture diagram, demo, Loom, GitHub, screenshots, or case study
Confidential information does not need to be shared. Anonymized evidence is acceptable.
Not a Good Fit If Your Experience Is Mainly
- Basic ChatGPT integrations
- Simple OpenAI API integrations
- Basic RAG chatbots
- Prompt engineering
- Zapier / Make automation
- AI content-generation workflows
- Simple AI assistants
- Tutorial projects
- Academic projects
- POC-only agent systems
Our internal team already handles these areas.
We need an architect capable of taking the system to the level of a secure, scalable, production-grade AI management platform.
What We Expect From You
We expect you to challenge our assumptions and provide strong architectural recommendations.
You should be able to explain:
- What architecture would you recommend and why?
- When should agents be used versus deterministic workflows?
- How should agents communicate?
- How should memory and business knowledge be managed?
- How should live business data be accessed securely?
- How should autonomous actions be controlled?
- How should human approval work?
- How should AI decisions be evaluated?
- How should prompt injection and unauthorized actions be prevented?
- How should agent failures and conflicting information be handled?
- How should the platform scale as the number of agents and business functions increases?
Pay: ₹40,000.00 - ₹60,000.00 per month
Application Question(s):
- Have you personally designed and deployed an Agentic AI or Multi-Agent AI system in production?
- How would you architect an AI Manager that can plan tasks, delegate work to multiple agents, review outputs, access live business data, request human approval for risky actions, execute approved API actions, and verify results?
- Have you designed security controls for AI agents that can access business systems, databases, or APIs?
- Which of the following have you used for production Agentic AI systems?
LangGraph
LangChain
MCP
Custom agent orchestration
OpenAI Agents / equivalent
Other
None
- Have you implemented Human-in-the-Loop approval workflows for AI actions?
- Have you implemented evaluation, monitoring, tracing, or observability for AI agent workflows?
- Have you built AI agents that access live business data and execute actions through APIs/tools?
Location:
- Mohali, Punjab (Required)
Work Location: In person
