AI products · Platform · DevOps

Yadlapalli Sankar Sai Narayana

Building intelligent systems
that ship reliably.

View work Résumé Get in touch
Focus LLM apps & APIs

Agents, RAG, evals, and product UX around models.

Stack Cloud & delivery

Kubernetes, CI/CD, IaC, and SRE-minded operations.

Principle Ship with confidence

Monitoring, testing, and guardrails from day one.

About

I am Yadlapalli Sankar Sai Narayana, an AI and Cloud Software Engineer focused on building reliable, production-ready AI systems. My work spans LLM applications, RAG architectures, Kubernetes platforms, observability, and cloud-native infrastructure.

At Yellow.ai, I work on AWS EKS production systems, lead incident response, and develop AI-powered observability tools that help engineers diagnose and resolve issues faster. As a Certified Kubernetes Administrator (CKA) with a B.Tech (Honors) in Information Technology, I am passionate about building scalable, resilient systems that deliver measurable impact.

  • 1+ yr Production AI & platform
  • CKA Kubernetes certified
  • ~40% MTTR improvement (on-call)
  • 11 MCP tools · unified diagnostics

AI & LLM systems

  • LangChain and LangGraph orchestration for agents and tool use
  • RAG pipelines, vector search, and Hugging Face embeddings
  • MCP integrations and autonomous RCA for distributed traces
  • Authoring custom AI agent skills and diagnostic tooling
  • Real-time STT/TTS pipelines and live speech translation

Software & APIs

  • FastAPI, Flask, and REST APIs for model-backed services
  • FastMCP and Python/TypeScript for integration layers
  • PostgreSQL and MySQL for application and analytics data

Cloud & DevOps

  • AWS EKS, Docker, Helm, and GitHub Actions CI/CD
  • On-call SRE, IAM/RBAC, and multi-environment cluster operations
  • Distributed tracing, log analysis, and incident management
  • Real-time telephony debugging: SIP routing and call signal tracing

Education & Work Experience

  1. Experience

    SDE (Infra, DevOps, Cloud, AI & Voice Stack) — Yellow.ai

    Mar 2025 — Present

    Yellow AI (Yellow Messenger), Bangalore — CKA-certified engineer owning AWS EKS operations and an MCP-powered observability platform used daily by engineering.

    Cloud infrastructure & SRE

    • Managed EKS across 3+ environments; upgrades and patching cut deployment failures by ~30%.
    • On-call for 15+ critical incidents; MTTR down ~40% via structured log analysis and remediation.
    • Right-sized workloads for ~25% lower infra overhead; CI/CD on GitHub Actions + Docker accelerated releases ~35%.
    • IAM, cluster RBAC, and routine patching across AWS components.

    AI platform & observability

    • Unified 11 MCP tools into one on-demand diagnostic platform; incident investigation from ~60 min to under 5.
    • LLM layer routes natural-language queries (trace ID, region) to the right tool — ~3 hours/day saved on manual debugging.
    • LangChain orchestration for autonomous RCA on distributed traces; MTTD down ~45%, developer productivity up ~50%.
  2. Education

    Certified Kubernetes Administrator (CKA)

    2025

    Linux Foundation, USA — production cluster admin, troubleshooting, networking, storage, and security for cloud-native workloads.

    • Directly supports EKS ownership at Yellow.ai: upgrades, RBAC, and multi-environment reliability.
  3. Education

    B.Tech Information Technology (Honors)

    Nov 2021 — May 2025

    JNTU-GV College of Engineering, Vizianagaram — GPA 8.0+ / 10. Honors graduate with focus on software engineering, AI/ML, and systems.

    • Coursework: algorithms, databases, networks, OS, and cloud-oriented subjects.
    • Labs & projects: TensorFlow, PyTorch, Hugging Face, Jupyter, Power BI, and ETL-style analytics pipelines.

Expertise

AI, LLM & Agents

Orchestration, RAG, and tool-use systems in production.

LangChain LangGraph MCP / FastMCP RAG pipelines Hugging Face Agentic workflows Custom AI agent skill authoring Claude Codex

Cloud & Kubernetes

EKS operations, cluster lifecycle, and AWS foundations.

AWS EKS EC2 / ECR IAM & RBAC Helm CKA Multi-env clusters

DevOps & CI/CD

Repeatable delivery, containerization, and Linux ops.

Docker GitHub Actions CI/CD pipelines Linux Git Release automation

Backend & APIs

Services that expose models, data, and integrations.

Python FastAPI Flask REST APIs PostgreSQL MySQL Node.js SQLite / Turso

Observability & SRE

On-call, tracing, and incident response at scale.

Distributed tracing Incident management On-call SRE Log analysis Cluster monitoring Autonomous RCA OpenSearch

Platform & Architecture

Cloud-native design for reliable, scalable systems.

Microservices Cloud-native System design Self-healing systems Infrastructure planning

Real-time & Telephony Systems

SIP call routing, live speech pipelines, and signal-level debugging.

SIP / VoIP routing Kamailio Real-time media streaming Speech-to-text (STT) Text-to-speech (TTS) Real-time speech translation Call signal tracing Multi-service call flow debugging OpenSearch analytics Call signal tracing Multi-service call flow debugging

Work

MCP · Platform · LLM 2026

TicketOps

Production agentic system at Yellow.ai: FastMCP + LangChain routes queries to 11 observability tools — cutting incident investigation from ~60 minutes to under 5.

Internal production tool GitHub profile
RAG 2026

SE-HE-RAG

Developed a self-healing RAG pipeline with adaptive retrieval and claim verification. The system automatically retries searches and regenerates responses when evidence is insufficient, ensuring reliable and grounded answers.

Live Demo Repository
Go · Python · K8s 2026

KubeAtlas

AI-powered Kubernetes operations platform that monitors cluster health, analyzes incidents, and identifies root causes in real time — combining observability, LLM-driven investigation, and automated remediation workflows.

In active development Repository
BI · Analytics 2024

Sales insights & e-commerce analytics

End-to-end analysis with MySQL, Power BI, and Jupyter — cleaning, modeling, and dashboards for business KPIs.

Sales insights E-commerce analysis
RAG · Local LLM 2025

Local RAG knowledge assistant

Privacy-first RAG assistant that runs entirely on-device — local embeddings, vector search, and LLM inference over your own documents, with no data leaving the machine.

Repository

Let’s connect

Open to AI engineering, platform, and DevOps roles. Based in Bangalore — reach me by email.

Email me LinkedIn GitHub