Software Engineer

Sai Phanindra

Platform Engineering - Backend Systems

Open to conversations

India

Learning:AI Infra & eBPF

I build backend systems and the platforms around them. My work spans internal developer platforms, AI agent infrastructure, security automation, and developer tooling. I enjoy turning complex engineering workflows into simple developer experiences.

Email me

Stack

Backend: Python, Go, TypeScript, Node.js, FastAPI, Gin, gRPC
Platform: AWS, Kubernetes, CI/CD, Docker, Terraform, Helm, Backstage
Data: PostgreSQL, Apache Kafka, MongoDB, Redis, RabbitMQ, ClickHouse, Snowflake
Observability: OpenTelemetry, Splunk, New Relic, Prometheus, Grafana

Work

Celigo

Developed an AI Agent SDK and platform with MCP, A2A, auth, skills/tools, evaluations, observability, and cost tracking, enabling teams to build and operate production-ready agents.

Developed MCP servers and integrations for internal engineering systems, enabling AI agents to securely interact with developer tools, services, and engineering workflows.

Built an internal developer platform with a golden path from GitHub through Jenkins, ArgoCD, and Kubernetes, enabling teams to provision and deploy services without platform tickets.

Built self-service local development infrastructure for 250+ engineers to spin up LocalStack and Pub/Sub across 70+ microservices, with automated environment setup and E2E testing.

Built a dependency security platform using OSV across 250+ services and 14 shared Node.js, Rust, and Python packages, with Jira integration, secure-version recommendations, and automated vulnerability remediation.

Built engineering productivity and integration platforms that aggregate data from internal systems and automate workflows, including API documentation-to-integration scaffolding used across 100+ connector releases.

During a monolith-to-microservices migration, correlated infrastructure spikes with Splunk telemetry to identify and diagnose 45 production issues.

Latest Writing

All writing

Aug 23, 2026 10 min

Read

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

SessionVault is a hosted MCP server that lets Claude and Cursor share the same session context. Encrypted at rest, tenant-isolated by Postgres RLS, searchable by pgvector.

Feb 28, 2026