Lead Backend Engineer
Ernst & Young (EY) / Bangalore
- Lead backend and API development for Ask Research, an enterprise GenAI research assistant built with Java 21, Spring Boot, Python/FastAPI services, and Kafka.
- Integrate OpenAI models to deliver grounded conversational answers over curated research data, refining prompts and retrieval context for answer quality and latency.
- Develop a retrieval-augmented generation (RAG) pipeline using vector embeddings and semantic search, supporting 90+ requests per minute.
- Expose backend capabilities to language models through the Model Context Protocol (MCP) and tool calling, enabling discovery and invocation of enterprise services.
- Prototype multi-agent workflows in which specialised agents coordinate retrieval, tool calls, and task hand-offs.
- Evaluate Amazon Bedrock and Google Vertex AI for model access, guardrails, and deployment trade-offs.
- Improve API latency by approximately 40% through caching, asynchronous processing, JVM tuning, and query optimisation.
- Apply Hexagonal Architecture to keep core business logic decoupled from model providers, vector stores, and data sources.
Java 21Spring BootPythonFastAPIKafkaMongoDBAWSRAGVector SearchMCPLLMs