EvalOps Control Plane
ongoingA control plane for managing prompt versions, offline evaluations, and release gates for LLM-backed features.
next.jstypescriptnode.jsclickhouse
RAKSHAN HEGDE
Software Engineer | AI Builder
Production-grade software engineer building AI-native products, agent workflows, and reliable systems.
terminal://rakshan-intro
█
Featured Projects
Production-facing work spanning agent systems, retrieval, and AI reliability.
Agent Observability Platform
ongoingA production-focused observability layer for multi-agent workflows with traces, eval snapshots, and latency/cost diagnostics.
next.jstypescriptpostgresopenailanggraph
RAG Knowledge Orchestrator
completedA retrieval pipeline that combines hybrid search, chunk-level reranking, and source-grounded response synthesis.
pythonfastapitypescriptpostgresqdrant
AI Signals
Hands-on capabilities I use to build resilient AI products.
- #LLM agents with tool orchestration
- #Retrieval-augmented generation (RAG) systems
- #Prompt/version control and evaluation pipelines
- #On-device and local inference experiments
- #Latency/cost-aware AI productionization