About

About CaL

CaL is a Toronto AI lab building software and teaching people to use AI. The work has agreed limits, and a person approves every release.

About us

Ali runs Citadel Agentic Labs in Toronto. The lab builds software and teaches people to use AI, with written limits on what agents may do and a person responsible for every release.

For more than ten years Ali has worked inside operations he did not build: trading floors, banks, and now city halls and small businesses. He builds ways of working that hold up after he leaves the room. CaL is that work, run as its own lab.

Founder and operator

Ali Akber, CFA

  1. Graduated in Business
  2. CFA Charter
  3. Trading floor
  4. KPMG
  5. Queen's
  6. Citadel Agentic Labs

Our Responsible AI framework

We agree what AI may do, check its sources and use only the data the task needs. A named person reviews the work and remains accountable for each decision.

We bake human review into the process. A named person approves what ships.

CaL follows ISO 42001 principles and makes no certification claim.

  1. Scope

    We agree what the AI works on before it starts.

    Write down the task, its limits and the decision it supports before using a model.

    Written approval records what the AI may do and which data it may use.

  2. Permissions

    It only gets access to what the task needs.

    Give people and systems only the access needed for the approved task.

    Actions with real consequences need approval from the person responsible.

  4. Release-and-monitor

    A person approves the release and keeps watching it.

    A named person approves release, checks the result, and responds when evidence changes.

    Record the approval and monitor the services the released system depends on.

See the guardrails

Alliances