About the Role
We are looking for an Agentic AI Engineer to design and build autonomous, multi-agent AI solutions for clients, preferably within the Consumer Packaged Goods (CPG), Food & Beverage industry. This is a hands-on engineering role for someone who enjoys turning technical requirements into working agentic systems — writing the orchestration logic, integrating tools/APIs, and iterating on agent reliability and performance.
You'll work closely with an Agentic AI Lead/architect to implement solution designs, and directly with client technical teams to understand system requirements and constraints.
Key Responsibilities
Requirements Understanding
- Work with the Agentic AI Lead, solution architects, and client technical stakeholders to understand functional and technical requirements for agent-based use cases.
- Review existing client systems, APIs, and data sources to understand integration points and constraints for agent tool-use.
- Clarify ambiguous requirements by asking the right technical questions and validating assumptions early through prototypes.
Agentic Solution Development
- Design and build multi-agent workflows using frameworks such as LangGraph, AutoGen, CrewAI, or Semantic Kernel, based on the solution architecture defined by the lead.
- Implement agent roles, task planning/decomposition logic, and coordination patterns (sequential, hierarchical, parallel) as designed.
- Build tool-use and function-calling integrations connecting agents to enterprise systems (ERP, SCM, CRM, internal APIs) and data sources relevant to CPG operations.
- Develop and tune prompts/instructions for agent reasoning, task execution, and inter-agent communication.
- Implement retrieval-augmented components (RAG, vector search) and structured data lookups to ground agent outputs in accurate information.
Testing, Evaluation & Reliability
- Build evaluation harnesses to test agent task completion rates, reasoning accuracy, and failure modes.
- Debug and resolve issues in agent behavior — infinite loops, incorrect tool calls, hallucinated outputs, or coordination failures.
- Implement guardrails, fallback logic, and human-in-the-loop checkpoints as specified in the solution design.
- Monitor agent performance in staging/production and iterate based on real usage patterns and feedback.
Deployment & Collaboration
- Package and deploy agentic solutions in collaboration with DevOps/platform engineering teams, following established CI/CD and infrastructure practices.
- Instrument agents with logging and observability (e.g., LangSmith, Arize, or custom tooling) to support debugging and performance tracking.
- Document agent designs, prompts, tool schemas, and known limitations for maintainability and knowledge transfer.
- Participate in code reviews and contribute to reusable components, templates, and best practices for the team.
Required Qualifications
- 3–6 years of relevant experience in software engineering, AI/ML engineering, or Gen AI development, with hands-on experience building agentic AI / multi-agent systems.
- Practical experience with at least one agent orchestration framework (LangGraph, AutoGen, CrewAI, Semantic Kernel, or equivalent).
- Strong Python programming skills; experience working with LLM APIs (OpenAI, Anthropic, Azure OpenAI, etc.).
- Experience building tool-use/function-calling integrations connecting LLMs/agents to external APIs and enterprise systems.
- Working knowledge of RAG pipelines and vector databases (Pinecone, Weaviate, Milvus, pgvector, or similar).
- Understanding of prompt engineering fundamentals and iterative prompt/instruction tuning.
- Comfortable working with technical requirements handed down from an architect/lead and asking clarifying questions to fill gaps.
- Familiarity with version control (Git), CI/CD basics, and collaborative software development practices.
- Exposure to or interest in CPG / Food & Beverage industry use cases preferred (e.g., supply chain, demand planning, trade promotion, customer service automation).
- Bachelor's degree in Computer Science, Engineering, Data Science, or a related field (or equivalent practical experience).
Preferred Qualifications
- Experience with agent evaluation/observability tooling (LangSmith, Arize, or similar).
- Familiarity with cloud platforms (AWS, Azure) and containerization (Docker/Kubernetes).
- Understanding of responsible AI practices — guardrails, bias mitigation, and safe failure handling for autonomous systems.
- Prior experience in a client-facing consulting or delivery environment.
Pay: ₹474,863.22 - ₹1,776,636.14 per year
Work Location: Hybrid remote in Delhi, Delhi (Delhi)