About
About
Practical engineering notes from production AI systems.
I am Kiran Scaria, a Principal AI Engineer focused on building dependable AI systems that ship and hold up in production.
Most of my work sits at the intersection of agent frameworks, memory systems, observability, and evaluation. I use this blog as an engineering journal: what worked, what failed, and what I would improve next.
If you are new here, start with the latest posts for current implementation notes.
Selected Open Source Work
- Raga LLM Hub: Evaluation tooling for LLM pipelines.
- AgentNeo: Python SDK for tracing, debugging, and observing agentic systems.
- RagaAI Catalyst: Hosted observability and evaluation workflows for production AI applications.
What You Will Find Here
- practical deep dives on agent architecture and memory patterns
- implementation notes on LLM infrastructure and reliability
- project retrospectives from systems built in the open
- tooling and workflows that improve engineering feedback loops
Outside software, I spend time with photography. You can see that work on my photo blog.
If this writing is useful, connect through the links in the footer.