Building at the edge of possibility.
Rohit's AI engineering lab, in public. Deep-dive writing, 90+ free browser tools, daily AI news, honest tool reviews, and live experiments across AI agents, Tesla telemetry, and self-hosted systems.
Explore the lab
writing, tools, news, and live experiments - all in one place
Perplexity open-sources Lily inference engine; Qwen releases local search layer
Perplexity released Lily, a Rust-Metal inference engine for Apple Silicon showing 1.35x decode speedup over MLX-LM; Qwen open-sourced zg, a unified local-first search combining ripgrep, BM25, and.
Blog
Deep dives on agentic AI, local LLMs, and self-hosted infra.
Open →Apps
Browser-based calculators and dev tools. No signup, no ads.
Open →News
AI industry briefings, neutral and cited. Catch up in a minute.
Open →Stack
Honest reviews of the tools actually in the workflow.
Open →Experiments
Engineering case studies and lab notes, post-mortem honest.
Open →Recent Posts
thoughts on engineering & AI
500 Days and 30,000 Miles
Five hundred days ago I pointed a Raspberry Pi at my car and started keeping every mile it drove. It has captured 10.4 million data points without missing a beat I did not cause, and somewhere along the way it stopped being a logger and became something I can ask questions of.
The AI Watermarking Arms Race
Anthropic has started watermarking Claude's text output. How generative text watermarking actually works, why no third party can verify it yet, why detection power drops to the square root of the machine-written fraction once a document has two authors, and why I shipped a tool that says 'not checkable' instead of inventing a score.
Introducing Gekro News: An AI Briefing That Curates Itself Around Me
A public daily AI briefing that reads a profile of my interests distilled from my own knowledge base, picks the day's real signal, cites its sources, and publishes itself every morning.
Latest Experiments
technical case studies & lab notes
Neural Home Orchestrator
A localized agentic system for managing home automation and energy efficiency. This experiment integrates local LLMs with Home Assistant to create a truly private, intelligent home controller.
Self-Hosted LLM Lab
Running Llama 3 on a local Raspberry Pi cluster with Ollama and Cloudflare Tunnels. This experiment explores the limits of edge compute for decentralized intelligence.
EV Data Logger
Transitioned from a custom TypeScript listener to a self-hosted TeslaMate fork running on a Raspberry Pi 5. Tracks every drive, charge, and phantom drain on my Model Y with 100% data sovereignty.