Sai Phanindra

Backend & Platform Engineer

India

Learning:AI Infra & eBPF

I build self-service platforms and distributed systems that make software delivery safer and faster. My recent work supports 70+ microservices and 250+ engineers, spanning golden paths, CI/CD automation, reliability, and AI-assisted developer tooling.

Get in touch

Technology Stack

Languages & Backend

Python

TypeScript

Node.js

FastAPI

Gin

Cloud & Platform

AWS

Docker

Kubernetes

Terraform

CI/CD

Backstage

Data & Messaging

PostgreSQL

MongoDB

Redis

Apache Kafka

RabbitMQ

ClickHouse

Snowflake

Observability & Reliability

Splunk

New Relic

OpenTelemetry

Experience

Celigo

Engineered a self-service internal developer platform with golden paths from GitHub bootstrapping through Jenkins, ArgoCD, and Kubernetes deployment.

Built a Python automation engine for 70+ microservices, giving 250+ engineers self-service LocalStack and Pub/Sub environments for shift-left validation.

Architected an anomaly-detection pipeline for a monolith-to-microservices migration, uncovering 45+ production issues by correlating infrastructure spikes with Splunk logs.

Built security automation for 250+ services and 14 shared packages, using OSV audits to generate vulnerability-remediation pull requests across Node.js, Rust, and Python.

Created a self-service AI workflow that converts API documentation into integration scaffolding, supporting release cycles of 100+ application connectors.

Engineered a self-healing browser automation engine that adapts selectors to DOM changes at runtime, maintaining 99.9% resilience for data-ingestion pipelines.

Expanded distributed execution coverage from 10% to 70% by integrating event-driven validation directly into the core CI/CD pipeline.

Developed an LLM-assisted automation workflow for Slack and Zendesk, reducing manual incident-triage time by 40% through classification and event routing.

Stabilized legacy distributed test suites and removed bottlenecks in internal automation frameworks, reducing build failures and pipeline flakiness.

Latest Writing

FULL ARCHIVE

Mar 7, 2026 6 min

Read

How I gave my AI tools a shared memory using MCP and pgvector

How I built SessionVault with MCP, PostgreSQL, and pgvector so AI tools can share decisions, files, and context without manual copy-paste.

Feb 28, 2026

WebSockets in Production: Heartbeats, Reconnects, and Backpressure

Feb 22, 2026

Streaming Like ChatGPT: Build Server-Sent Events from Scratch

Featured Projects

VIEW ALL

Open Source

pg_vitals

Modern terminal UI for live PostgreSQL performance analysis. Monitors multiple databases at once with real-time views for activity, slow queries, locks, indexes, memory, and safe EXPLAIN plans. Built on Textual and async psycopg 3 with per-target connection pools and background refresh workers for a responsive UI.

PythonTextualpsycopg 3+5

View Project

Open Source

SafeStream

AI-assisted YouTube live chat moderation stack. Monitors live chat in real time, applies customizable moderation rules using pluggable AI backends (OpenAI, Gemini, LM Studio), and provides a secure dashboard for managing sessions. Designed for local, Electron, or server deployment with strong security defaults.

Node.jsElectronLangGraph+3

View Project

Live · Code

techdigest

A minimal open-source RSS reader for following engineering publications in one place. It aggregates posts from multiple sources into a fast, focused reading experience without recommendation algorithms or feed noise.

Next.jsTypeScript

Live Demo