Need MCA/B.Tech/ Candidate only

Role/Job Description

Experience - 5-8 Years

Budget - 80,000/Monthly

Key Responsibilities
- RAG Pipelines: Design and implement end-to-end Retrieval-Augmented Generation systems — including chunking strategies, embedding models, vector stores, hybrid search, and re-ranking — to deliver accurate, context-grounded LLM responses.
- Agentic AI Development: Build autonomous and multi-agent AI workflows using frameworks such as LangChain, LangGraph, AutoGen, CrewAI, or Semantic Kernel; implement tool-use, planning, memory, and orchestration patterns.
- Knowledge Graphs: Model, build, and query knowledge graphs using Neo4j and other Graph Databases; integrate graph-based retrieval (GraphRAG) with LLM pipelines for enhanced reasoning and explainability.
- LLM Integration: Integrate and fine-tune Large Language Models (LLMs) using prompt engineering, function calling, structured outputs, and parameter-efficient techniques (LoRA/QLoRA) where applicable.
- Deployment & MLOps: Containerize and deploy GenAI services on AWS, Azure,



or GCP; implement monitoring, evaluation, versioning, and cost-efficient scaling for AI workloads.
- Responsible AI: Apply guardrails to mitigate hallucinations, prompt injection, bias, and data leakage; contribute to evaluation frameworks for model accuracy and safety.
- Collaboration: Partner with cross-functional teams, document technical designs clearly, and communicate trade-offs effectively with both technical and non-technical stakeholders.

Required Technical Skills
- Generative AI: Robust hands-on experience building GenAI applications using LLMs (OpenAI GPT, Anthropic Claude, Llama, Mistral, Gemini, etc.); solid grasp of Transformer architectures, embeddings, and prompt engineering.
- RAG: Proven experience designing RAG pipelines — chunking, embeddings, vector databases (Pinecone, Chroma, Weaviate, Milvus, FAISS, pgvector), hybrid search, and re-ranking.
- Agentic AI & Tools: Hands-on experience with Agenti

📌 Generative AI (India)
🏢 Sparix Global
📍 India