NY.

AI ENGINEERING / PRODUCTION SYSTEMS

Narendra Yechuri

From intelligence
to working systems.

AI Engineer

Generative AI · RAG · Production Backend Systems

I build production AI applications combining LLMs, retrieval, backend APIs, real-time systems, and cloud infrastructure. My recent work includes enterprise RAG platforms, real-time Voice AI systems, and agentic automation workflows.

View projects Download resume
GitHub ↗LinkedIn ↗Email ↗
Practical AI.Considered architecture.Production ownership.

01 / SELECTED WORK

Systems I’ve built.

From real-time conversations to grounded answers and orchestrated workflows.

01REAL-TIME AI Production

Real-Time Voice AI Platform

Senior Engineer / Lead Backend Developer

A conversational AI backend coordinating speech, retrieval, live API data, and LLM inference for concurrent telephonic conversations.

PythonPipecatGemini 2.5 FlashVertex AIPostgreSQL / pgvectorSoniox STTCartesia TTSTata Tele BusinessREST APIs

ENGINEERING NOTES

  • Designed asynchronous processing for concurrent conversations and latency-sensitive service coordination.
  • Used PostgreSQL and pgvector for semantic retrieval, with real-time API data fetched when required.
  • Implemented horizontal autoscaling. Kubernetes with Horizontal Pod Autoscaling is a planned future architecture.

CONCEPTUAL FLOW

  1. Caller
  2. Telephony
  3. STT
  4. AI backend
  5. Retrieval / APIs
  6. Gemini
  7. TTS
  8. Caller

High-level flow; implementation details omitted.

02RETRIEVAL & KNOWLEDGE Production

Enterprise Knowledge GenAI Platform

End-to-End Owner — Architecture, Development, Deployment & Support

An enterprise knowledge assistant grounding answers in organizational sources through Retrieval-Augmented Generation.

PythonFastAPIGPT-4ClaudeOpenAI EmbeddingsQdrantAzure AI SearchAzure Function AppsApp ServiceBlob StorageKey VaultPrompt Flow

ENGINEERING NOTES

  • Owned the microservices architecture, Python/FastAPI backend, Azure deployment, and production support.
  • Built semantic and vector retrieval with OpenAI Embeddings, Qdrant, and Azure AI Search to supply relevant context to GPT-4 and Claude.
  • Implemented LLM evaluation and observability with Prompt Flow and custom telemetry for latency, data validation, and prompt changes.

CONCEPTUAL FLOW

  1. User
  2. FastAPI
  3. Query / embedding
  4. Vector search
  5. Context
  6. GPT-4 / Claude
  7. Response

High-level flow; implementation details omitted.

03AGENTIC WORKFLOWS Proof of Concept

AI DevOps Automation Agent

Workflow design & development

An exploration of agentic architecture for coordinating CI/CD and deployment-related tasks through multi-step tool workflows.

PythonLangGraphMCPn8nCI/CD concepts

ENGINEERING NOTES

  • Designed multi-step agent workflows and tool orchestration.
  • Explored MCP integrations and n8n automation for CI/CD and deployment workflows.

Professional projects are described at a conceptual level. Employer source code and private implementation details are not published.

02 / EXPERIENCE

Ownership across the lifecycle.

Dec 2021 — Aug 2026

Cerebra Consulting Inc

Software Engineer

Owned architecture, backend development, cloud deployment, and production support for enterprise AI and data platforms. Built a knowledge assistant with OpenAI GPT, Claude, Qdrant, and Azure AI Search; PDF/OCR extraction; and Azure Functions-compatible APIs. Developed time-series recommendations with XGBoost and LightGBM, migrated APIs from SQL Server to Azure Cosmos DB, and built database synchronization and diagnostic tooling. Additional work included an AI DevOps Automation Agent POC and computer vision POCs using TensorFlow and PyTorch.

03 / TOOLKIT

The tools behind the systems.

Organized by the problems they help me solve.

Generative AI

  • GPT-4 / OpenAI
  • Claude
  • Gemini / Vertex AI
  • Prompt Engineering
  • RAG
  • Embeddings
  • Semantic Search
  • LLM Evaluation

Agentic AI

  • LangGraph
  • MCP
  • n8n

Backend

  • Python
  • FastAPI
  • REST APIs
  • Microservices
  • Async / Concurrent Processing

Vector & Data

  • Qdrant
  • PostgreSQL
  • pgvector
  • Azure AI Search
  • SQL Server
  • Azure Cosmos DB
  • MongoDB

Cloud & Operations

  • Azure Function Apps
  • Azure App Service
  • Azure Blob Storage
  • Azure Key Vault
  • Azure AI Foundry / ML tooling
  • GCP
  • Vertex AI
  • Terraform
  • CI/CD
  • Prompt Flow

Real-Time AI

  • Pipecat
  • Soniox
  • Cartesia
  • Telephony integrations

04 / ABOUT

Practical by design.

I’m an AI and Software Engineer with a background spanning software, AI, and engineering. I’m interested in practical production AI: systems that work beyond the demo.

I take end-to-end ownership, from requirements and architecture through backend development, deployment, and ongoing support.

EDUCATION & LEARNING

Master of Technology in Nanotechnology

Jawaharlal Nehru Technological University, Kakinada
2018–2022

Bachelor of Technology in Mechanical Engineering

Sanketika Institute of Technology & Management, JNTU-Kakinada
2017

Certification

Artificial Intelligence and Machine Learning

THE FULL PICTURE

Experience, in one document.

A concise overview of my projects, engineering experience, and skills.

Download resume

05 / CONTACT

Let’s talk about
what you’re building.

For AI engineering roles and conversations about production AI systems.

GitHub ↗LinkedIn ↗