Rajesh Medampudi

Engineering leader with 17+ years building production observability platforms, real-time distributed systems, and scalable gaming backends.

rajesh@medampudi.com · LinkedIn · GitHub · Hyderabad, India

Experience

Technical Manager

Head Digital Works

Apr 2023 – Present

MLT Platform — Enterprise Observability

Built and operate a Grafana LGTM stack (Loki, Mimir, Tempo, Alloy) on AWS EKS, ingesting 6 TB/day across 15 departments at $15K/month — 85-90% below commercial SaaS pricing.

  • Karpenter autoscaling with 10+ specialized node pools on ARM Graviton2 — 20-40% cheaper than x86
  • NVMe instance store caching with 80% hit rate, S3 Glacier lifecycle policies saving 96% on storage
  • Blue-green deployments across dual production clusters, 99.5% platform uptime
  • SLO dashboards and Zenduty alerting for 200+ developers
Loki · Mimir · Tempo · Grafana · Karpenter · AWS EKS · S3 · Helm

RSTS — Real-Time Segmentation & Targeting

Designed a real-time player targeting engine processing millions of events daily at <500ms p99 latency using Java 21 virtual threads, Cassandra, and RabbitMQ.

  • 40+ parallel Cassandra queries per event in <200ms using prepared statement caching
  • JEXL expression engine for zero-code campaign rules — delivery improved from 60% to 99.99%
  • CQRS architecture separating admin writes from high-throughput query path
  • React admin portal with real-time DAG visualization
Java 21 · Spring Boot 3.2 · Cassandra · RabbitMQ · Hazelcast · React · Redux

Engineering Leadership

  • Lead 15-member team across platform infrastructure, gaming backends, and mobile
  • Mentored 2 engineers into team lead roles managing 7-8 reportees each
  • Drove engineering certification culture — team earned CKAD, Node.js, Go, React certs

Assistant Manager

Head Digital Works

Jan 2022 – Mar 2023

Casual Games Platform — Real-Time Multiplayer Backend

Architected Gen 3 distributed gaming backend serving 100K+ concurrent players at 2,000 TPS across 4 game titles.

  • Custom Netty WebSocket server with Hazelcast distributed state — leveraged ALB sticky sessions for connection affinity
  • 3-tier microservice architecture: Lobby → Matchmaking → Game Server
  • Optimized AWS spend through right-sizing, Karpenter, Graviton, and Reserved Instances
  • 30+ mobile app releases shipped with service-oriented design patterns
Java 17 · Spring Boot · Netty · Hazelcast · RabbitMQ · AWS ECS · DynamoDB

Independent Projects

Self-hosted Infrastructure

2019 – Present

Building and operating production infrastructure on bare metal with Kubernetes, zero-trust networking, and full observability.

  • Edge security gateway: HAProxy + CrowdSec + WAF with automated SSL
  • Zero-trust mesh: Tailscale connecting all servers, no public IPs on backends
  • Self-hosted LGTM observability across 7+ servers
  • Vertical SaaS applications on Frappe/ERPNext — 22+ features shipped

Key Numbers

6 TB/day Telemetry ingestion
$120K/mo AWS cost savings
100K+ Concurrent players
<500ms p99 event latency
15 Engineers managed
17+ yrs Engineering experience

Technical Skills

Cloud & Infrastructure

AWS (EKS, ECS, S3, DynamoDB, IAM), Kubernetes, Karpenter, Helm, Terraform, Docker

Observability

Grafana, Loki, Mimir, Tempo, Alloy, Prometheus, OpenTelemetry, Zenduty

Backend

Java 17/21, Spring Boot, Python, Go, Node.js, Netty, RabbitMQ

Data

Cassandra, PostgreSQL, DynamoDB, Hazelcast, Redis, S3

Frontend

React, TypeScript, Redux Toolkit, Astro, Tailwind CSS

Self-Hosting

K3s, Talos, HAProxy, CrowdSec, Tailscale, Let's Encrypt, Frappe