Yash Krishan

I'm a backend engineer with 5+ years of experience designing and scaling production systems from high-performance APIs and automation pipelines to LLM-powered agent infrastructure.

I’ve worked across the stack, but my core focus is on building robust backend services that strike a balance between performance, reliability, and maintainability. At Haystack Analytics, I’ve led projects ranging from traditional data workflows to modern AI systems, including multi-agent coordination, RAG pipelines, and streaming infrastructure.

My current work bridges backend engineering with LLM infra: using Python, Rust, DuckDB, MongoDB, and frameworks like Agno and the Model Context Protocol (MCP) to build memory-aware, real-time agent workflows.

I'm especially interested in:

- Building scalable, modular backend systems

- Making LLM-based workflows production-grade

- Architecting infrastructure for data-intensive and AI-driven applications

- Simplifying dev workflows through thoughtful tooling

Currently exploring: orchestration patterns for AI agents, streaming-first UI backends, and fast analytics with DuckDB.