Responsibilities:

Gen AI · Architect and implement GenAI solutions using LLMs (e.g., GPT, Claude, Gemini) for tasks like summarization, code generation, document intelligence, and conversational AI · Fine-tune and optimize models using frameworks like Hugging Face, LangChain, and OpenAI APIs · Integrate GenAI capabilities into enterprise platforms (CRM, ERP, knowledge bases) Agentic AI Development · Design and deploy autonomous AI agents capable of multi-step reasoning, tool use, and goal-driven execution · Build agentic workflows using frameworks such as AutoGen, CrewAI, LangGraph, and Semantic Kernel · Enable agents to interact with APIs, databases, and external tools securely and reliably Cloud & Infrastructure · Develop scalable AI pipelines on cloud platforms (Azure, AWS, GCP) using Kubernetes, serverless, and containerized environments · Implement MLOps practices for continuous integration, monitoring, and governance of AI systems Client Engagement · Collaborate with global stakeholders to define AI use cases, success metrics, and delivery roadmaps · Lead offshore delivery teams and ensure alignment with client expectations and timelines · Contribute to reusable assets, accelerators, and internal capability building

Mandatory skill sets:

· 3+ years of experience in AI/ML engineering, with 2+ years in GenAI and/or Agentic AI · Proficiency in Python, PyTorch/TensorFlow, and GenAI/Agentic frameworks · Hands-on experience with cloud-native AI services (Azure ML, AWS SageMaker, GCP Vertex AI) · Familiarity with vector databases (e.g., FAISS, Pinecone), RAG pipelines, and prompt engineering · Strong understanding of autonomous agents, orchestration, and multi-agent systems · Excellent communication and stakeholder management skills across geographies

Preferred skill sets: · Certifications in cloud platforms and AI technologies · Exposure to enterprise-grade AI use cases in BFSI, healthcare, retail, or manufacturing · Knowledge of Responsible AI, data governance, and compliance frameworks

Years of experience required: