# Everett Kleven Engineer and founder. I build AI-native simulation substrates — the layer between a model deciding something and that decision having consequences. Creator of Archetype (ECS simulation engine on Pydantic, Daft, and Iceberg) at Vangelis Technologies. Adjacent interest: Physical AI, robotics, and agent execution you can replay and audit. - Email: ekleven@vangelis.tech - GitHub: https://github.com/everettVT - X: https://x.com/everettkleven - LinkedIn: https://www.linkedin.com/in/everett-kleven - Company: https://vangelis.tech ## Talks ### Claude Code in the Wild: 3 Builder Demos (+ Fable 5) - Date: 2026-06-28 - Venue: Rally SF - Series: Building the State of the Art - Role: Co-host - Recording: https://www.youtube.com/watch?v=yqKZo6kmjrs Three builders demo what they actually ship with, plus a first look at Fable 5. Orchestration and harness design carried the night: how much context an agent needs to act usefully, and how to keep the loop verifiable once it does. ### Claude Code and Beyond: State of the Art Agentic Coding Workflows - Date: 2026-06-23 - Venue: Rally SF - Series: Building the State of the Art - Role: Co-host - Recording: https://www.youtube.com/watch?v=H2JQR1-oAhg The fourth installment, and the longest. Skills dominated — how builders package repeatable procedure so an agent can be trusted with it — alongside review discipline, evals, and the first real comparisons across Claude Code, Codex, and Fable. ### Claude Code and Beyond: 4 Real Agentic Coding Workflows - Date: 2026-06-07 - Venue: Rally SF - Series: Building the State of the Art - Role: Co-host - Recording: https://www.youtube.com/watch?v=1ChmzWNLAK0 Four builders walk through the workflow they genuinely use, end to end. The recurring threads: writing specs an agent can execute, keeping pull requests small enough that a human can still review them, and where token budgets stop being a detail and start dictating architecture. ### How 2 Engineers Rebuilt a $2M Platform in 2 Months with AI Agents - Date: 2026-04-13 - Venue: Rally SF - Series: Building the State of the Art - Role: Co-host - Recording: https://www.youtube.com/watch?v=wbjQS_eqEco The series' first full case study: two engineers replace a $2M platform in two months of agent-driven work. Concrete on the parts that usually get waved at — how the work was specified, split into reviewable pull requests, and gated on tests rather than vibes. ### Fever Dreams: Working With and Against AI Sycophancy - Date: 2025-07-24 - Venue: Nerds Like Me — Chicago - Role: Speaker Why an assistant's agreeableness is a failure mode rather than a courtesy. The talk walks the good (the barrier to shipping software has collapsed), the bad (human-feedback finetuning that rewards flattery over truth), and the ugly (the GPT-4o sycophancy episode and so-called "ChatGPT psychosis") — then lands on Narcissus as the mental model: a reflection you cannot pull away from, even as it costs you. Position: LLMs are mirrors, not oracles, and every output is arbitrary until vetted. Structured as Definitions → The Good → The Bad → The Ugly → Mental Model → Closing. Draws on Sharma et al. on sycophancy in language models, the ELEPHANT framework for face-preserving behaviours, Anthropic's work on reward tampering ("Sycophancy to Subterfuge"), and DeepMind on synthetic data as a mitigation. The closing is the part worth keeping: outputs are arbitrary, so truth is approached across many experiments rather than accepted from one; and long-term self-isolation with a model that agrees with you is the red flag to watch for. ## Projects ### Archetype - Role: Creator - Status: Active - Source: https://github.com/VangelisTech/archetype - Stack: Python, Daft, Iceberg, Pydantic, ECS An AI-native ECS simulation engine: Pydantic's composability for typed components, Daft dataframes for distributed processing, and Iceberg for multi-modal storage. Entities are rows, systems are transformations, and world state is a table you can query after the fact. Archetype treats simulation state as columnar data rather than object graphs. That makes a world's history queryable — you can reconstruct what happened, when a transition fired, and which validator gated it, instead of inferring it from logs. The same properties that make it useful for simulation make it useful as an execution substrate for agent work: typed task graphs, durable checkpoints, restore after failure, and evidence that survives the run. ### world-examples - Role: Author - Status: Active - Source: https://github.com/VangelisTech/world-examples - Stack: Python, Archetype, examples Minimal component, prefab, system, and graph composition examples for Archetype — the smallest working thing for each concept, so the engine can be learned by reading one file at a time. ### awesome-physical-ai - Role: Maintainer - Status: Active - Source: https://github.com/everettVT/awesome-physical-ai - Stack: robotics, embodied AI, curation A curated list of open-source projects for Physical AI — robotics, embodied intelligence, simulation, and the infrastructure that makes it work. ### sessionscope - Role: Creator - Status: Active - Source: https://github.com/everettVT/sessionscope - Stack: HTML, local-first, agent tooling See how you actually work. A privacy-first browser analyzer for Claude Code, Codex, and ChatGPT session files — zero upload, runs entirely offline. ### gepa - Role: Contributor - Status: Exploring - Source: https://github.com/everettVT/gepa - Stack: Jupyter, optimization, prompting Optimizing prompts, code, and more with reflective text evolution — treating the artifact under improvement as something a model can critique and rewrite against a measurable objective.