Posted on: 31/08/2025
Role Overview :
- Develop and optimize RAG (Retrieval-Augmented Generation) systems using efficient chunking strategies for context-aware responses.
- Use LangChain to orchestrate modular GenAI workflows integrated with vector databases (e.g., FAISS, Pinecone, Chroma).
- Build and maintain knowledge graphs and incorporate Vision APIs for multimodal intelligence.
- Apply advanced prompt engineering and token management to control costs and improve LLM accuracy.
- Ensure reliability and robustness through hallucination control methodologies and responsible AI practices.
- Design and deploy scalable GenAI applications on cloud platforms (Azure, AWS, or GCP).
- Collaborate with cross-functional teams including data scientists, cloud engineers, and domain experts to integrate GenAI into products.
- Leverage tools like Docker for containerization and Git for code versioning.
- Use SQL and Python to process and manipulate data effectively.
Must-Have Skills :
- Retrieval-Augmented Generation (RAG)
- Chunking Strategies for document processing
- Deep understanding of LLMs (e.g., OpenAI, Anthropic, Meta, Cohere)
- Experience with AI services on Azure, AWS, or GCP (at least one)
- LangChain framework
- Vector Databases (e.g., FAISS, Pinecone, Weaviate)
- Token Management strategies
- Knowledge Graphs (e.g., Neo4j, RDF-based models)
- Vision APIs (e.g., Google Vision, AWS Rekognition, Azure Computer Vision)
- Prompt Engineering expertise
- Proficiency in Python
Good-to-Have Skills :
- Deep Learning frameworks (e.g., PyTorch, TensorFlow)
- Computer Vision models and techniques
- Hallucination Control Techniques
- Responsible AI frameworks and compliance methodologies
Technical Proficiencies :
- Cloud : Azure, AWS, or GCP (at least one is mandatory)
- Containers : Docker
- Database : SQL
- Version Control : Git or similar
Qualifications :
- Relevant certifications in AI/ML or Cloud Platforms are a plus.
- Strong analytical thinking, problem-solving, and communication skills.
Did you find something suspicious?