Nasir Sotoudeh · Builder
I design and ship systems that move data from raw streams into reliable, intelligent products — combining strong engineering fundamentals with pragmatic AI tooling.
Turning a simple personal site into a small AI platform — a place to ship agent skills, data products, and long-lived experiments.
Direction
From batch ETL to always-on, AI-augmented systems.
Theme
Building systems, not just code — with clear constraints.
I'm Nasir Sotoudeh, a data engineer by training who's steadily moved closer to the product and platform layer. My background spans building and maintaining data pipelines, analytics workflows, and systems that teams can trust.
Over the last few years I've become increasingly focused on AI-native products — the glue between infrastructure, data, and user-facing experiences. I care about reliability, clarity, and making complex systems feel understandable.
This site is my laboratory for that direction: a small, opinionated platform where I can experiment with agentic workflows, data products, and the tooling around them.
In progress
A local system that turns long documents into structured, searchable knowledge.
It helps users quickly find answers and insights without needing technical interaction.
Runs fully on-device with fast, efficient retrieval and clear source-based results.
Coming soon
Experiments around data observability, feedback loops, and agent-assisted workflows for analytics and platform teams.
I write about systems, data platforms, and the messy reality of shipping AI-powered features. Expect fewer hot takes and more grounded, practical notes.
Medium
Longer-form breakdowns of systems, migrations, and the trade-offs behind specific architecture decisions.
Substack
Short notes on experiments, agent skills, and the "building in public" side of this platform.