Rahul N R.

Multi-agent AI system with seven domain-specialized agents

A local Python-based multi-agent framework where each agent is mapped to an anime character and handles a distinct life domain — fitness, finance, career, learning, wellbeing, planning, and confidence. Built for extensibility and open-source release.

Multi-agent AI framework with a deployed Telegram fitness coach

A Python-based multi-agent system where seven domain-specialized agents handle fitness, finance, career, learning, wellbeing, planning, and confidence. Toji — the fitness agent — is fully deployed on Telegram via Railway, using Groq for fast inference and ChromaDB for persistent memory across sessions.

AIOps observability platform with self-healing infrastructure

Transforms distributed systems into transparent, self-healing infrastructure using Prometheus and Grafana. Detects anomalies in real-time log streams and triggers automated recovery — bringing AI into the operational layer.

Local-first personal AI agent with permanent memory and hardened persona

The first agent of the Renkai empire. A privacy-first executive proxy running entirely on local hardware via Ollama, with dual-model routing — Llama 3.1 for reasoning, Qwen 2.5 Coder for code generation. Features Yggdrasil, a ChromaDB-backed long-term memory system that recalls past context automatically. Includes CI/CD via GitHub Actions, a full test suite, and a hardened persona engine — no chatbot apologies, just execution.

Agentic coding system with a floating AI assistant that plans, builds, and remembers

An intelligent agentic coding system featuring Lynn — a full-stack AI builder with three interfaces: a CLI agent for quick tasks, a Streamlit studio for full project context, and a floating PyQt5 desktop companion (think Clippy, but actually useful). Lynn clarifies requirements before touching a keyboard, plans across the stack, writes clean code, validates her own output, and remembers past sessions via vector memory.