Writing
Thinking out loud.
Engineering deep dives, career lessons, books, and whatever else I'm working through.
[ img ]
Building a real time LLM enrichment pipeline on Kafka
How I wired LLM calls into a live Kafka stream without blowing the latency budget, with a reliability layer and token control.
Read article →
→
[ img ]
Why pgvector over a dedicated vector database
The case for keeping your vectors in Postgres until you genuinely outgrow it, with the tradeoffs laid out.
Read article →
→
© 2026 Ramakrishna Sayee Meka built in the open