3 months ago
Bengaluru, IndiaStaff+
Responsibilities
- Lead the design and development of hybrid vector and graph retrieval architectures.
- Architect data pipelines for ingestion, embedding, and indexing of massive multimodal datasets.
- Prototype advanced retrieval techniques including reranking, LLM graph tooling, and metadata filtering.
- Design schemas for knowledge graphs and ensure high-performance relationship mapping and ontological integrity.
- Build data validation and drift-detection systems for embedding quality and vector-space health.
- Implement persistent, observable, low-latency memory systems for AI agents.
- Unify disparate data sources into coherent, searchable knowledge bases and establish data organization standards.
- Collaborate with AI Research and Product teams to evaluate and integrate emerging database and retrieval technologies.
Requirements
- 7+ years of experience in data engineering or backend systems focused on high-performance data retrieval and storage.
- Expert proficiency in Python, Java, or Go and strong knowledge of distributed-system design patterns.
- Deep understanding of vector-database indexing strategies including HNSW, IVFFlat, and PQ, along with cosine, Euclidean, and dot-product distance metrics.
- Experience with Pinecone, Milvus, Weaviate, or Qdrant.
- Strong experience with graph databases such as Neo4j, AWS Neptune, or ArangoDB and query languages including Cypher or Gremlin.
- Experience building semantic layers, ontologies, taxonomies, and other data models.
- Hands-on experience with LangChain, LlamaIndex, and embedding models from OpenAI, HuggingFace, or Cohere.
- Proficiency with Spark, Flink, or Kafka for large-scale processing, real-time indexing, and ETL.
- Understanding of information-retrieval concepts including BM25, TF-IDF, and reciprocal rank fusion.
- Experience with AWS, GCP, or Azure and Kubernetes.
- Bachelor's or master's degree in computer science or a related field, with 7–9 years of professional experience preferred.
Tech Stack
Categories
BackendData Engineering
About Hotstar
We’ve moved! 🌟 Follow us on @JioHotstar for all the latest stories, updates & more.
