Hermes Agent
privateSelf-hosted AI agent with a Telegram gateway, running rootless Docker under systemd.
- python
- docker
- linux hardening
- systemd
case study →
$ whoami
Senior Software Engineer
Amogh Shendre
Building reliable AI agents and mobile systems.
Nine years shipping production software at Fortune 500 scale — currently leading React Native for a retailer whose app reaches millions. I now apply that same rigor to tool-using agents, retrieval systems, mobile platforms, and the edge hardware I run in my homelab.
01selected work
all projects →RAG Infra-Docs Chat
privateChat with my infrastructure docs — hybrid retrieval over pgvector, evaluated with promptfoo.
- postgres
- pgvector
- hnsw
- hybrid search
- promptfoo
case study →
BLE Biosensor Pipeline
privateESP32-S3 heart-rate monitor with gap-detection and backfill sync to Android and a Pi backend.
- c++
- esp32-s3
- max30102
- kotlin
- node.js
case study →
02latest writing
all posts →15 min read
Self-hosting Qwen on a headless Mac mini
How I run Qwen with Ollama: on a MacBook for trying models, and on an always-on Mac mini behind Tailscale as a private model server for a real app. The headless-Mac gotchas, and the cold-load bug that made every quiz question take 158 seconds.
- homelab
8 min read
Letting an AI agent run my Pi-hole maintenance
Two days of overdue homelab maintenance (251 pending packages, a Pi-hole upgrade, community regex rules, a broken Grafana Alloy install), done by Claude Code over SSH, with sudo kept firmly in human hands.
- homelab
- agentic-ai
10 min read
Building a voice agent that drives with me
How I turned a headless mini PC and an existing AI agent into a hands-free voice assistant for the car, and the two bugs that taught me the most.
- agentic-ai
- homelab