AI Enabled Software Development — engineered, not assembled.
STELEIOS is a software and data engineering practice based in Jaipur, working with clients across India's IT hubs and beyond. We build systems that carry real operational load — data pipelines, streaming platforms, and the applications that sit on top of them.
Every engagement is AI-enabled: machine learning, language models and vector retrieval are part of how we build, not a feature bolted on at the end.
We favour boring, durable choices: fewer dependencies, explicit contracts between layers, and tests that cover the edge cases rather than only the happy path. Security and cost are design inputs from the first commit, not something retrofitted once the system is live.
Our writing reflects the same standard — the blog publishes first-hand build notes from systems that were actually run, including the parts that broke.
Go — compiled services with a small memory footprint and predictable latency.
TypeScript end to end — typed state management and server-side rendering, so a page arrives complete for readers and machines alike, on one of the fastest JavaScript runtimes.
PostgreSQL, using JSONB for nested documents and pgvector for semantic search.
Python with Polars and DuckDB — columnar, embedded, and fast enough to run beside the application.
Batch and incremental data pipelines — extraction, warehousing, orchestration and testing, built on open-source tooling such as Airflow, dbt and DuckDB.
Real-time ingestion and processing for high-volume streaming and IoT workloads, designed to behave correctly under backpressure.
Assessment and modernisation of existing systems, reducing operational IT cost while improving reliability.
Machine learning, deep learning and large language model capabilities built into production software rather than bolted on.
Design and operation of cloud and distributed systems, with an emphasis on cost efficiency and operational simplicity.