The hands-on academy from /tmp Labs. We turn your teams into practitioners who can build, secure, govern, and certify AI for regulated, real-world deployment — graded on production-grade labs, taught by the people who operate these systems under scrutiny.
Most AI training teaches concepts. This builds capability you can audit — from the same practitioners who run AI systems under regulatory scrutiny in banking, insurance, and healthcare.
The curriculum comes from operating production systems that must be explainable, auditable, and accountable — not from a textbook.
Labs mirror /tmp Labs’ own governance and security systems — the techniques are the ones we actually run, not toy examples.
Every module is anchored to verified incidents and live regulation — EU AI Act, MAS, NIST AI RMF, ISO/IEC 42001 — so the lessons are current and real.
Each module ends in an auto-graded gate with anti-gaming checks. The capability is demonstrated, not asserted — the standard an auditor applies.
Every week is a real engagement, not a lecture. Learners build a defense, discover a trap, and prove the fix holds — against checks they can’t game.
A tight, incident-led briefing — the why before the how.
A starter repo with a seeded flaw. Build the real defense, on a real stack.
A flaw learners discover for themselves — not one they’re told about.
An auto-grader proves the fix works and holds on held-out variants.
Start anywhere that fits the role. Build literacy, then practitioner depth, then specialize — with a parallel track for the leaders who govern it all.
Literacy and governance framing for the people who sign off on AI risk.
The full ladder is running now — from first-step awareness to flagship specialist certifications. Everything is hands-on and gated, and available at the altitude you need (see delivery).
Literacy, obligations, and the questions to ask — for boards, CISOs, CROs, and heads of AI.
Certified AI judgment for students — catch AI mistakes, verify what you’re told, and use AI with integrity. No coding. Built for any stream, and the graduates employers trust.
Broad, practical AI fluency for every role in the organization — the common on-ramp.
Hands-on core skills for engineers building with AI — from data to deployment, with labs.
Senior-level engineering depth, delivered as stackable modules so teams take only what they need.
Build, secure, and govern agentic systems and tool use — hands-on, for teams shipping agents now.
Operate AI like critical infrastructure — lineage, CI/CD gates, monitoring, MRM evidence packs. For platform, MLOps & risk teams.
See an AI system as an adversary does, then make it safe — threat modeling to red-team capstone. For security & AI red teams.
Defend when the most capable AI is on both sides — machine-speed adversaries, and your own AI in the line of fire. From the AI-transformed threat landscape to a live blue-team capstone. For CISOs, SOC & blue teams.
We’ll map a pathway to your stack, your risks, and your team’s roles in a short scoping call.
Map my pathway →The same rigorous content, packaged for how your people can actually learn — from a quick self-paced on-ramp to a full certification.
Scales across the whole organization. A guided cohort cadence keeps completion high.
The essential 20%, fast. Ideal for teams that need capability now, or an exec wedge.
The full depth, instructor-led, with a credential. The premium, capacity-limited tier.
Foundations, Advanced, Agentic — build AI that holds up in production.
Security for AI — attack, defend, monitor, and certify AI systems.
Cyber Defense in the Age of AI — detect, hunt, defend, and respond at machine speed.
AI Ops & Governance — evidence packs and assurance auditors will accept.
Governance & Risk for Leaders — the literacy to govern AI with confidence.
Run one real week against your own environment — see the labs, the gate, and the artifacts before you commit. Then roll out the pathway that fits your teams.
For individuals and self-paced learners, Karpavai is the /tmp Labs Academy learning platform for practical AI courses, submissions, simulations, and verifiable credentials.