@harsh017

Harsha

@harsh017Mumbai, IndiaJoined May 2026

The future of CI/CD is already running in production.

About

Passionate about AI, DevOps, and building scalable developer-focused solutions.

Available for

Nothing here yet.

Harsha's blogs

How to Build an AI Agent from Scratch with LangChain and OpenAIlangchainandopenai.hashnode.dev10 posts

Articles Comments1

Recently published

HHarshalangchainandopenai.hashnode.devAug 17 · 10 min read

Your AI Isn't Slow. Your Legacy Architecture Is.

They need better infrastructure. Companies are spending heavily on AI models, agents, copilots, and automation while some of the systems feeding those AI products still depend on batch processing, dis

HHarshalangchainandopenai.hashnode.devAug 14 · 14 min read

Enterprise AI Is an Engineering Problem, Not a Model-Selection Problem

There is a strange pattern happening in enterprise AI. A company builds an impressive prototype. The model produces surprisingly good answers. Executives see the demo. Everyone gets excited. Then the

HHarshalangchainandopenai.hashnode.devAug 13 · 8 min read

Traditional Architecture Wasn't Built for AI , And We're Pretending It Was

Most engineering teams know how to build scalable web applications. We have microservices. REST APIs. Kubernetes. Autoscaling. CI/CD. Observability. Load balancers. We've spent years turning these pat

HHarshalangchainandopenai.hashnode.devJul 30 · 4 min read

No-Code Is Great for Validation. But Serious Products Eventually Need Real Engineering.

No-code platforms deserve a lot more credit than they get. They help founders validate ideas quickly, launch MVPs in weeks, and avoid spending months building software nobody wants. But here's the opi

HHarshalangchainandopenai.hashnode.devJul 16 · 4 min read

Stop Designing the Same UI Twice: Why Code-to-Design Sync Is the Future of Product Development

If there's one workflow in modern product development that I think deserves to disappear, it's this: Developer builds a feature → Designer recreates it in Figma → Both versions drift apart → Everyone