Skip to content
View ancelin-code's full-sized avatar
🧠
Teaching agents to remember, one drawer at a time
🧠
Teaching agents to remember, one drawer at a time
  • Schneider Electric ⚡
  • Grenoble, France
  • 11:16 (UTC +02:00)
  • LinkedIn in/ancelinserre

Highlights

  • Pro

Block or report ancelin-code

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
ancelin-code/README.md

Hello world, I'm Ancelin 👋

AI Engineer at Schneider Electric ⚡

My day job is AI and data products for the energy sector (Python, Azure, Databricks). On the side, I build the tooling for the AI-native workflow I use every day: specs refined with agents, parallel lanes in git worktrees, an independent agent reviewing every change, tests and ADRs as the source of truth.

Open to remote roles in AI engineering, developer productivity or AI-native developer tooling. French (native), English (professional).

Projects

  • delivery-kit: Agent Skills that take a project from requirements to reviewed pull requests, in parallel lanes, with git as the only state. Works with Claude Code, Copilot CLI, Codex, OpenCode and Pi. POSIX sh helpers tested on Linux, macOS and Alpine, decisions recorded as ADRs.
  • tiroir: configurable, scoped memory for AI agents. Plain Markdown in Git, strict isolation between drawers, local hybrid search, served over MCP.

Stack

Python · FastAPI · Databricks · MLflow · Azure · Terraform · OpenTelemetry · MCP · Agent Skills · Coding agents (Claude Code, Copilot CLI, Codex, OpenCode)

Pinned Loading

  1. delivery-kit delivery-kit Public

    Agent Skills that take a project from requirements to reviewed pull requests, in parallel lanes, with git as the only state. Works with Claude Code, Copilot CLI, Codex, OpenCode and Pi.

    Python

  2. tiroir tiroir Public

    Configurable, scoped memory for AI agents: plain Markdown in Git, user-defined drawers, strict isolation, local hybrid search, MCP.

    Python