Aakash Ojha

I’m a Software Engineer focused on enterprise AI infrastructure and production-grade LLM applications. I independently design backend tools and documented solutions that cut developer workflow effort by 90%, while translating business and developer needs into working systems.

In my current role as a Software Engineer Trainee, I built an enterprise AI assistant that retrieves knowledge across SAP EWM, Jira, Confluence, and SharePoint through a unified conversational interface. I designed an ingestion layer with SAP RFCs, OData Services, CDS Views, and SAP HANA, and engineered a RAG platform indexing ~100 GB of enterprise knowledge with ~25 million vector embeddings using Qdrant and Azure AI Search.

I improve retrieval quality with hybrid search (BM25, BM42, dense retrieval), BGE-M3 embeddings, BGE reranking, semantic ranking, and document-aware semantic chunking, orchestrated through 30+ LangGraph nodes for multi-source retrieval and clarification workflows. I also ship production APIs and integrations using 16 FastAPI REST endpoints, containerize services with Docker, and deploy through Azure DevOps CI/CD with unit and integration tests for reliability.