Job Description :

We are looking for a Lead ML Engineer to take full ownership of the customer next-generation search platform. You will bridge advanced information retrieval (IR) modeling with high-throughput production systems — building, tuning, and scaling hybrid search pipelines while owning the evaluation harness, performance benchmarks, and master data integrations that guarantee extreme precision and sub-second latency at scale.
This is a high-impact, end-to-end engineering role. You will shape the architecture of a system that directly affects the experience of millions of customers, while mentoring talented engineers and partnering closely with product and data stakeholders.
WHAT YOU'LL DO
Search Core & Information Retrieval
Design, build, and tune a hybrid search engine combining Vector Search (semantic) and Lexical Search (BM25/keyword) to deliver best-in-class relevance.
Implement advanced ranking and blending strategies including BGE rerankers and Reciprocal Rank Fusion (RRF).
Own end-to-end search accuracy and relevance metrics (NDCG, MRR, Recall) and drive measurable,



data-backed improvements over time.
Evaluation & Quality Assurance
Build and maintain an automated accuracy evaluation harness for continuous, regression-proof pipeline testing.
Establish quality benchmarks and champion a metrics-first engineering culture across the team.
Performance, Scalability & Infrastructure
Conduct systematic load testing (Locust, k6) and stress-test retrieval pipelines to surface and eliminate bottlenecks.
Architect and optimize systems to guarantee a strict SLA of P95 latency < 500ms under peak production load.
Partner with DevOps/MLOps to design scalable, resilient deployment patterns for search and ranking models.
Data Engineering & Governance
Manage ingestion pipelines for Master Data Management (MDM) feed integration, ensuring clean and timely data synchronisation.
Govern schema and operational configurations within the Firestore spec_registry.
Collaborate

📌 Lead-ML Engineer (Hyderabad)
🏢 Egen
📍 Hyderabad