A Sai Bharath

A Sai Bharath

AI Backend Engineer at NCR Atleos, Hyderabad

I design the backend behind LLM products: pipelines that make sense of messy enterprise data, search that finds the right source, and agents that only run when a question actually needs one.

Open to AI backend and LLM platform roles

Projects

AI analytics platform

Freelance, private

A natural-language analytics product for a client: business users ask questions about their company data in plain English and get answers back. Client work, so the code and details are private.

Python, FastAPI, LangGraph, Azure OpenAI

Dockwave

Live

Document Q&A across a whole workspace. Entities are extracted into a Neo4j knowledge graph so answers can follow connections between documents that vector search alone misses. Several retrieval passes are graded and the most confident answer wins.

Next.js, FastAPI, LangChain, Neo4j, PostgreSQL, Azure OpenAI, S3

PhisX

56 users on the Chrome Web Store

A Chrome extension that flags phishing pages while you browse, backed by an ML model with 95%+ accuracy served from a REST API on EC2.

Python, Flask, React, AWS EC2, Docker

Personal Investor

In development

A finance assistant that reads credit-card statements from Gmail and lets an LLM choose the right finance tool for each question, so questions don't need a fixed format.

Next.js, Python, LangChain, Gmail API

Experience

Oct 2025 – present

Associate Software Engineer

NCR Atleos, Hyderabad

  • Building internal AI and automation tooling on production LLM and retrieval services.

Feb 2025 – Oct 2025

Software Developer Intern

NCR Atleos, Hyderabad

  • Built a Graph RAG pipeline (LangChain + Neo4j) for multi-step Q&A over internal knowledge bases.
  • Built LLM services on Azure OpenAI with prompt pipelines, retrieval validation and failure handling.
  • Set up GitHub Actions CI/CD that removed manual deploy steps and kept dev and staging in sync.

Skills

Retrieval and agents
RAG, hybrid search (BM25 + HNSW), RRF reranking, Graph RAG, tool calling, evaluation and guardrails
Backend
Python, FastAPI (async), LangGraph, LangChain, Celery, Redis, Pydantic, SQLAlchemy
Data
PostgreSQL + pgvector, OpenSearch, Neo4j, MongoDB, DuckDB, DataFusion, Parquet
Cloud and frontend
Azure OpenAI, AWS (EC2, S3), Docker, GitHub Actions, TypeScript, React, Next.js

Hiring for an AI backend role?

Send me the role and a little context, and I'll reply by email with what I'd bring to it.