Skip to content

Igor Fil

I'm Igor, an AWS Solutions Architect helping startups build resilient, scalable cloud systems. With nearly a decade building the cloud at Amazon and in the cloud for startups to enterprises, I design solutions that handle spikes, cut costs, and drive growth. I advise select startups, turning tech challenges into wins.

Everyone Is Afraid of Missing a Train That Isn't Coming

Recruiters flood my inbox 3-4 times a week with Forward Deployed Engineer offers. Postings are up 700%. Billions are being spent on a job that didn’t exist two years ago. I’ve seen this stampede before — microservices, Kubernetes, Kafka — and I keep saying no.

RAG Is Just a Search Engine Bolted to a Language Model

RAG — Retrieval Augmented Generation — is a search engine bolted to a language model. Instead of stuffing all your documents into the prompt (impossible) or baking facts into weights via fine-tuning (inflexible), RAG retrieves only the relevant chunks at query time and feeds them to the model. This post breaks down how indexing and query pipelines work, why hybrid search and reranking matter, and when to use RAG vs. fine-tuning.