👋

Hi, I am Barun Debnath

I build full-stack products, AI systems, and the platforms that keep them reliable, observable, and scalable when real users show up.

My lane sits where product engineering meets SRE: TypeScript and Go services, agent workflows, ClickHouse/Kafka data paths, Kubernetes platforms, and the reliability work behind them.

Open To Opportunities

infra
ai
ml
agents
sre
scaling

Current Arc

Full-stack x AI

Shipping product experiences, agents, APIs, and the infra they depend on.

Main Class

Platform-minded builder

Go, TypeScript, ClickHouse, Kafka, Kubernetes, and production debugging.

Side Quest

Writing Field Notes

Turning systems, AI, and hard-earned reliability lessons into posts.

💼 What I've Been Up To

Selected work across full-stack engineering, AI systems, platform reliability, and scalable infrastructure.

Founding Engineer LinkRunner Nov'25 - Present

Rebuilt real-time attribution pipelines around Go, Kafka, ClickHouse, and Redis for high-volume event ingestion.
Delivered major infra cost reductions across analytics, Kubernetes workloads, logging, and cloud migration work.
Led platform reliability work across GKE, GitOps, ClickHouse operations, Redis recovery paths, and compliance evidence.

GoClickHouseKafkaGKERedisCost

SRE/Platform Engineer One2N Jun'24 - Nov'25

Built hybrid cloud and Kubernetes platform work with EKS, Cilium, Istio, CoreDNS, ExternalDNS, and ingress routing.
Reduced GenAI inference cost by moving workloads from managed APIs to self-hosted model serving on SageMaker.
Scaled and hardened multi-cloud infrastructure with VPCs, load balancers, WAF rules, VPNs, CI/CD, and ISO/SOC2 work.

EKSCiliumIstioSageMakerTerraformSecurity

SRE/DevOps Media.Net Jul'23 - Jun'24

Built monitoring surfaces for production queueing pipelines with Django, Prometheus metrics, and Grafana dashboards.
Handled zero-downtime GKE node-pool upgrades, production MySQL migration work, and Redis spot-node cost optimization.
Improved registry reliability and cluster cost attribution with Artifact Registry migration and Stackdriver exporter work.

GKEPrometheusGrafanaMySQLRedisDjango

SRE Intern Media.Net Jan'23 - Jun'23

Built internal Terraform modules to standardize resource provisioning across environments.
Migrated GitOps deployments from ArgoCD app-of-apps to ApplicationSet patterns for better environment scale.
Integrated DriftCTL with Terraform and Atlantis to detect and manage infrastructure state drift.

TerraformArgoCDApplicationSetAtlantisDriftCTL

From product surfaces and AI workflows to production reliability and scalable platforms, this is the operating range. Know more about my work experience from my RESUME

Lore

Know more about me

Origin Class

Full-stack engineer with platform instincts

I build product surfaces, backend systems, AI workflows, and the infra paths that keep them reliable when traffic, data, and debugging pressure increase.

Current focus Full-stack + AI

Reliability lens SRE / Platform

Favorite terrain Systems at scale

Proof of Work

Shipping where product and infrastructure meet

Recent work spans real-time attribution, ClickHouse analytics, GenAI cost optimization, Kubernetes platforms, observability, and zero-downtime cloud migrations.

Events handled 25M+ / day

Infra cost cut 25x

Inference cost cut 99%

Traffic scale 10x growth

Backpack

Skills I carry into the build

A practical stack for building apps, agents, platforms, and production systems end to end.

Languages

GoPythonTypeScriptJavaScriptRustSQLC++

Backend & APIs

GoNode.jsFastAPIExpressgRPCRESTDjangoFlask

AI Engineering

RAGAgentsLangGraphLangChainMem0TemporalMCPOpenAI API

Data

ClickHouseKafkaRedisPostgreSQLMongoDBNeo4jQdrantElasticsearch

Cloud & Platform

GCPAWSKubernetesDockerCiliumIstioCloudflare

Reliability

PrometheusGrafanaThanosOpenTelemetryLokiAlertmanagerOpenCost

IaC & Delivery

TerraformTerragruntAnsibleArgoCDGitHub ActionsGitLab CIHelm

GitHub Realm

Open-source trail

A live-ish view of the public work and contribution trail I keep around tools, platform experiments, AI projects, and developer workflow systems.

View GitHub profile

Public repos 74

Original repos 55

Followers 132

Public stars 33

PythonAstroRustGoTypeScriptHCL

GitHub contribution activity graph for d-cryptic

Speaking

Talk Delivered

A recent speaking win on scaling real-time attribution systems with ClickHouse.

ClickHouse Bangalore User Group

25x Cheaper Infrastructure, 8x Cheaper Pricing: Real-Time Attribution With ClickHouse as the Backbone

Real-time attribution systems are analytical workloads: high-volume event ingestion, time-window queries, and heavy aggregations. This session covers how we rebuilt Linkrunner's attribution system from PostgreSQL to ClickHouse, reduced infrastructure cost by roughly 25x, and unlocked pricing around 8x cheaper than competitors.

Event: Agentic AI Meets Real-Time Data: Building the Intelligence Layer of Tomorrow

When: Saturday, Apr 18, 2026, 10:00 AM - 2:00 PM IST

Where: AWFIS Residency Square, Bengaluru

Speakers: Darshil Rathod, Co-founder & CTO @ Linkrunner, and Barun Debnath, Founding Engineer @ Linkrunner

View Meetup event

Latest posts

See all posts

Recent projects

See all projects

💬

Want to Chat?

Whether it's about full-stack products, AI engineering, system design, or reliability at scale, I'm always up for a good conversation.

Grab a coffee, poke around, and don't forget to say hi! You can find me at:

Languages

Backend & APIs

AI Engineering

Data

Cloud & Platform

Reliability

IaC & Delivery

Deploying a Scalable NATS Cluster - Part 2: Hands-On Demo

Hitchhiker's Guide To Make SadServers Happy - Part 1

Python Tricks that LLMs won't teach you - Part 1

Enough C++ To Build Anything: Part 1

Deploying a Scalable NATS Cluster - Part 1: Core Architecture and Considerations

Neovim Saga Part 1: Getting Started

CCSentinel

PaladinAI

ChattGator