- Core Programming & Systems Skills
- Python (expert level) for ML, orchestration, and agent logic
- Strong understanding of async programming, concurrency, and task scheduling
- Foundations of Agentic AI
- Design and implementation of autonomous AI agents capable of:
o Multi-step reasoning and planning
o Goal decomposition and task orchestration
o Dynamic decision-making under uncertainty
- Experience with agent architectures:
o ReAct, Plan-and-Execute, Reflexive agents
o Hierarchical / multi-agent systems
o Tool-augmented and function-calling agents
- Understanding of stateful vs stateless agents and memory management
- Large Language Models (LLMs)
- Hands-on experience with LLMs (OpenAI, Azure OpenAI, Anthropic, open-source models)
- Prompt-engineering techniques for:
o Reasoning (Chain-of-Thought, Self-Reflection)
o Planning and critique loops
o Instruction following and tool use
o Few-shot and zero-shot prompting
o Model selection trade-offs (latency, cost, context length)
- Knowledge of fine-tuning / adapters (LoRA) is a plus
- Agent Frameworks & Tooling
- Practical experience with agent frameworks, such as:
o LangGraph / LangChain (agents, tools, memory)
o Semantic Kernel
o AutoGen, CrewAI, or similar
- Ability to build custom agent orchestration layers beyond frameworks
- Tool abstraction and execution safety (timeouts, retries, sandboxing)
Classification: Internal
- Memory, Context & Knowledge Augmentation
- Design of agent memory systems:
o Short-term (conversation/state memory)
o Long-term (episodic, semantic memory)
- Retrieval-Augmented Generation (RAG):
o Vector databases (FAISS, Pinecone, Azure AI Search, etc.)
o Embedding selection and chunking strategies
- Techniques for context management and compression
- Knowledge graph–augmented or hybrid memory (plus)
- Planning, Reasoning & Control
- Experience implementing:
o Task planners (step planning, re-planning)
o Constraint-based execution
o Feedback and self-correction loops
o Tool reliability scoring
o Guardrails and action validation
o Failure detection and graceful recovery
- MLOps & AgentOps
- Deployment of agents into production environments
- Observability for agents:
o Tracing agent decisions and tool calls
o Logging prompts, responses, and errors
- Model and prompt versioning
- CI/CD for agent systems
- Experience with Docker, Kubernetes, serverless deployments (Azure/AWS)
- Evaluation & Testing of Agentic Systems
- Designing evaluation frameworks for agents:
o Task success rate
o Cost, latency, and reliability
o Safety and hallucination detection
- Offline test harnesses and simulation environments
Classification: Internal
- A/B testing of prompts, tools, and agent strategies
- Security, Safety & Responsible AI
- Secure tool execution and privilege control
- Prompt-injection and jailbreak risk mitigation
- Data privacy and isolation in agent memory
- Responsible AI practices:
o Bias awareness
o Explainability of agent decisions
o Human-in-the-loop escalation patterns
10. Data & Integration Skills
o Enterprise systems (CRM, ERP, databases)
o Web services, internal APIs, and SaaS tools
o SQL / NoSQL databases
o Event-driven systems and message queues (plus)
11. Cloud & Platform Expertise
- Strong experience with at least one cloud platform:
o Azure (preferred for enterprise agentic AI), AWS, or GCP
- Managed AI services, identity & access, secrets management
- Cost optimization for LLM-driven systems
12. Bonus / Advanced Skills (Nice to Have)
- Multi-agent collaboration and negotiation
- Human-AI collaboration patterns (copilots, supervisors)
- Reinforcement learning for agent policy optimization
- Experience building enterprise copilots or autonomous workflows
Pay: ₹406,293.45 - ₹2,072,258.15 per year
Work Location: Remote
