Open to opportunities

Ekaansh Sawaria

@ekaanshsawaria

Message

I build production RAG pipelines, LLM APIs, and scalable document intelligence systems.

India

Message

What I'm looking for

I'm looking to build scalable backend and AI systems, especially RAG pipelines, LLM-powered products, retrieval systems, and developer tooling where I can keep shipping and learning.

I've built production-focused RAG pipelines, LLM-powered APIs, and document intelligence systems, with a focus on system design, efficient inference, and developer tooling.

At Accenture, I led a team of three interns to architect an enterprise-scale multi-agent RAG platform. I developed an agentic framework with multiple AI agents and six dynamically callable tools, and implemented hybrid retrieval using PageIndexer, BM25, dense vectors, Cross-Encoder reranking, and QuadRank.

For RExplain, I built an AI-powered GitHub repository analysis platform that reduced analysis latency by 95% by using the GitHub Tree API for selective retrieval. I also built its RAG pipeline, FastAPI backend, React frontend, PostgreSQL embedding cache, and Dockerized deployment.

I've also trained satellite image classifiers reaching 97.2% accuracy and researched mixed-precision LLM quantization approaches that reduced memory use by up to 40% with under 2% perplexity degradation.

Experience

Work history, roles, and key accomplishments

Associate Software Engineering Intern

Accenture

May 2026 - Jul 2026 (2 months)

Led a team of 3 interns in architecting an enterprise-scale multi-agent RAG platform, establishing system architecture, retrieval workflows, and agent orchestration strategy. Developed a Grok Build 0.1-based agentic framework with multiple AI agents and 6 dynamically callable tools, and implemented a hybrid retrieval engine with PageIndexer, BM25, dense vector retrieval, Cross-Encoder reranking, a