Role & responsibilities
Build and maintain connectors from enterprise systems (Salesforce, SharePoint, S3, SQL DBs) to the AI retrieval layer. Handle incremental sync and change detection.
Design and optimise document chunking strategies for different content types (PDFs, HTML, structured tables, multi-modal). Maintain preprocessing pipelines.
Select, fine-tune, and maintain embedding models (Titan Embeddings, penAI Ada, BGE, multilingual models). Benchmark quality on domain data.
Manage OpenSearch Serverless / Pinecone / pgvector clusters: indexing, refresh cycles, cost optimisation, namespace management.
Implement hybrid search (dense + sparse), re-ranking, query expansion, and retrieval evaluation. Own the precision/recall benchmarks.
Enforce document-level access controls in retrieval, ensure PII redaction pipelines,



and maintain data lineage for all AI-consumed datasets.

Preferred candidate profile
4+ years in data engineering (Python, PySpark, or SQL-heavy roles)
Hands-on experience building RAG pipelines: chunking, embedding, vector retrieval
Proficiency with at least one vector database: OpenSearch, Pinecone, Weaviate, or pgvector
Solid SQL and ETL pipeline experience; familiarity with Airflow or Glue
Understanding of embedding models when to use domain-specific vs general-purpose
Experience with enterprise document formats: PDF extraction, SharePoint, Confluence, structured

📌 Artificial Intelligence Engineer Mumbai
🏢 Godrej Properties
📍 Mumbai