Member of Technical Staff at Sully.ai. Projects, writing, and research on context engineering, multi-agent systems, and production AI.
A specification method and Agent Skill for launching autonomous runs that span hours, context windows, or parallel workers, with explicit success predicates, non-counting outcomes, evidence requirements, and adversarial review.
A Git-based personal operating system for AI agents: 80+ plain-text files, no database, ~87% fewer tokens per task through progressive disclosure.
Claude Code plugin: autonomous multi-agent copywriting loop. Learns your voice, self-critiques, and rewrites until acceptance criteria pass.
Fine-tuned an 8B model on Gertrude Stein's literary style for about $2 in compute. Built as a book-sft-pipeline skill that composes with four other skills from the Agent Skills repo.
Open-source collection of transferable Agent Skills for building production AI agent systems on Claude Code, Cursor, or custom stacks.
How we use paired interviewer/observer agents to extract tacit expert judgment and turn it into deployable personas.
Fine-tuned an 8B Llama that beat its 104B teacher by about 11% F1 on a memory-routing benchmark, via prompt distillation and RL on Tinker (the teacher still wins on hard multi-label cases).
An adaptive multi-agent system that extracts your literary DNA through conversation and generates actionable reading profiles.
Personalized food tour planner using LangChain DeepAgents, Google Maps API, and Tavily research.
Experiments across prompt engineering: creative writing with Kimi K2, production prompt architecture, comparing reasoning techniques (ToT, GoT, LoT), structured prompting tests, and a public dataset of 4,643 marketing prompts.
Multi-agent system that analyzes any GitHub repo and generates a repo-specific coding assessment using Google's A2A protocol. 2nd place, Google Cloud × AI Tinkerers Hackathon.
One-click social post → research report for LinkedIn and X. Capture a post, auto-shape a research query, search with reasoning, get a detailed report.
Autonomous research agent that scrapes URLs, qualifies enterprise AI case studies, and generates structured reports. Built in a weekend in response to an Arvid Kahl tweet.
One of the early AI GTM builders behind Rootly's Clay-featured AI Demand Generation Engineer motion: autonomous account research, lead scoring, incident-triggered outreach, and personalized outbound at startup scale.
Daily AI automation newsletter on no-code tools, step-by-step guides, and tool comparisons. Manually curated; paused in late 2023 when the workload became unsustainable.
Early AI budtender and cannabis assistant prototype, publicly described by the oz. as the world's first A.I. budtender. A fun project that tested vertical AI assistants, product recommendation, domain safety, and consumer education before AI budtenders became common.
AI career platform connecting students to AI internships via autonomous agents. Multi-agent scraping, DSPy matchmaking, a 2,000-member student community, and a weekly newsletter. Paused 2024 to join 99Ravens.