AI Agents course

You are inworkspace

Core path

Your path through this level

Progress saves in this browser and syncs after you sign in

Completed

0 of 4

Pick up where you left off

Train, then verify Isolation and routing Ship, observe, recover Experiment, then decide

Optional

Full module map

Use this if you want the shape of the level before you start

Show

Module 5.1Train, then verify

Fine tuning is worthwhile only when you can test the win and the risk.

Open

Prerequisites

Comfort with earlier modules in this track
Ability to explain trade-offs and risks without jargon

Outcomes

Explain train, then verify in your own words and apply it to a realistic scenario.
Fine tuning is worthwhile only when you can test the win and the risk.
Check the assumption "Training data is safe" and explain what changes if it is false.
Check the assumption "Evaluation includes regressions" and explain what changes if it is false.

Practice

Solve a complex scenario with explicit assumptions and constraints
Write one mitigation plan and one fallback plan

Artefact and failure modes

A concise design or governance brief that can be reviewed by a team
Memorisation. A tuned model can regurgitate training data. Test for it.
Behaviour drift. A change that improves one task can silently worsen another.

Module 5.2Isolation and routing

Enterprise agents succeed when tenants are isolated and requests are routed by risk and cost.

Open

Prerequisites

Comfort with earlier modules in this track
Ability to explain trade-offs and risks without jargon

Outcomes

Explain isolation and routing in your own words and apply it to a realistic scenario.
Enterprise agents succeed when tenants are isolated and requests are routed by risk and cost.
Check the assumption "Tenant data is isolated" and explain what changes if it is false.
Check the assumption "Routing rules are explicit" and explain what changes if it is false.

Practice

Solve a complex scenario with explicit assumptions and constraints
Write one mitigation plan and one fallback plan

Artefact and failure modes

A concise design or governance brief that can be reviewed by a team
Cross tenant leakage. One leaked prompt or log line is enough to break trust and create legal risk.
Cost spikes. If routing always chooses the biggest model, your bill becomes a surprise.

Module 5.3Ship, observe, recover

Production systems need packaging, monitoring, and a plan for failure.

Open

Prerequisites

Comfort with earlier modules in this track
Ability to explain trade-offs and risks without jargon

Outcomes

Explain ship, observe, recover in your own words and apply it to a realistic scenario.
Production systems need packaging, monitoring, and a plan for failure.
Check the assumption "Health checks are real" and explain what changes if it is false.
Check the assumption "Rollback is possible" and explain what changes if it is false.

Practice

Solve a complex scenario with explicit assumptions and constraints
Write one mitigation plan and one fallback plan

Artefact and failure modes

A concise design or governance brief that can be reviewed by a team
Silent degradation. The agent can be wrong slowly. Monitor quality signals, not only uptime.
Runaway cost. A broken loop can burn tokens and tool calls until you hit a budget limit.

Module 5.4Experiment, then decide

Treat research as a pipeline. Try ideas safely, evaluate honestly, then adopt deliberately.

Open

Prerequisites

Comfort with earlier modules in this track
Ability to explain trade-offs and risks without jargon

Outcomes

Explain experiment, then decide in your own words and apply it to a realistic scenario.
Treat research as a pipeline. Try ideas safely, evaluate honestly, then adopt deliberately.
Check the assumption "Experiments are safe" and explain what changes if it is false.
Check the assumption "Evaluation is comparable" and explain what changes if it is false.

Practice

Solve a complex scenario with explicit assumptions and constraints
Write one mitigation plan and one fallback plan

Artefact and failure modes

A concise design or governance brief that can be reviewed by a team
Hype adoption. If you adopt because it is popular, you inherit failure modes you did not choose.
No exit strategy. If a new technique fails, you need a path back to the last safe version.

What changes at this level

Level expectations

Each level is independent but clearly deeper than the last. This panel makes the jump explicit.

Learning contract

Assessment outcomes

Read the explanation first, then use the tools to test the idea. Skip any tool that is not useful for your goal.

AI Agents stage assessment. Advanced mastery

Timed stage assessment with free access. Sign in so attempts and pass evidence stay attached to your account.

Stage 5 of 6

Practice before your timed attempt

Review your stage notes and rerun one practical tool before starting.
Complete the matching practice assessment for this stage.
Sign in and start the timed attempt when your reasoning is stable.

See sample questions Practice first Review stage content Sign in

Assessments record pass evidence and need your name. Sign in to continue.

Before you start

First-try attempts that reach the top 20 may be invited to publish to a child-safe public leaderboard.

You choose whether to publish. If you decline, your top 20 record is not saved.

AI Agents stage assessment. Advanced mastery

This assessment is timed. Question count and pass threshold are shown for your session when it loads.

Enter your full name. This name is locked for your pass records and CPD evidence.

You can retake as many times as you like. Attempts are tracked. Recent questions are avoided when possible.

What to Expect

Question Types

Multiple choice (single answer)
Code analysis questions
Scenario-based reasoning
Security vulnerability identification

Topics Covered

Fine-tuning and LoRA
Enterprise architectures
Production deployment
Research frontiers

Assessment Rules

Rules

The exact time limit is shown when you start the attempt.
Questions are drawn from the live published bank. The exact question count is shown when your session loads.
The pass threshold is 70 percent for this stage.
You can retake and attempts are tracked.
AI assistants are disabled during timed sessions.
Copy and context menu actions are restricted to maintain assessment integrity.

CPD candidates: Use the CPD prep pack for structured revision before attempting. If you need help, email ransford.amponsah@ransfordsnotes.com with your CPD reference code.

CPD prep pack Track CPD

Course materials are protected by intellectual property rights.View terms

Loading lesson...

Core path

Your path through this level

Progress saves in this browser and syncs after you sign in

Completed

0 of 4

Pick up where you left off

Train, then verify Isolation and routing Ship, observe, recover Experiment, then decide

Optional

Full module map

Use this if you want the shape of the level before you start

Show

Module 5.1Train, then verify

Fine tuning is worthwhile only when you can test the win and the risk.

Open

Prerequisites

Comfort with earlier modules in this track
Ability to explain trade-offs and risks without jargon

Outcomes

Explain train, then verify in your own words and apply it to a realistic scenario.
Fine tuning is worthwhile only when you can test the win and the risk.
Check the assumption "Training data is safe" and explain what changes if it is false.
Check the assumption "Evaluation includes regressions" and explain what changes if it is false.

Practice

Solve a complex scenario with explicit assumptions and constraints
Write one mitigation plan and one fallback plan

Artefact and failure modes

A concise design or governance brief that can be reviewed by a team
Memorisation. A tuned model can regurgitate training data. Test for it.
Behaviour drift. A change that improves one task can silently worsen another.

Module 5.2Isolation and routing

Enterprise agents succeed when tenants are isolated and requests are routed by risk and cost.

Open

Prerequisites

Comfort with earlier modules in this track
Ability to explain trade-offs and risks without jargon

Outcomes

Explain isolation and routing in your own words and apply it to a realistic scenario.
Enterprise agents succeed when tenants are isolated and requests are routed by risk and cost.
Check the assumption "Tenant data is isolated" and explain what changes if it is false.
Check the assumption "Routing rules are explicit" and explain what changes if it is false.

Practice

Solve a complex scenario with explicit assumptions and constraints
Write one mitigation plan and one fallback plan

Artefact and failure modes

A concise design or governance brief that can be reviewed by a team
Cross tenant leakage. One leaked prompt or log line is enough to break trust and create legal risk.
Cost spikes. If routing always chooses the biggest model, your bill becomes a surprise.

Module 5.3Ship, observe, recover

Production systems need packaging, monitoring, and a plan for failure.

Open

Prerequisites

Comfort with earlier modules in this track
Ability to explain trade-offs and risks without jargon

Outcomes

Explain ship, observe, recover in your own words and apply it to a realistic scenario.
Production systems need packaging, monitoring, and a plan for failure.
Check the assumption "Health checks are real" and explain what changes if it is false.
Check the assumption "Rollback is possible" and explain what changes if it is false.

Practice

Solve a complex scenario with explicit assumptions and constraints
Write one mitigation plan and one fallback plan

Artefact and failure modes

A concise design or governance brief that can be reviewed by a team
Silent degradation. The agent can be wrong slowly. Monitor quality signals, not only uptime.
Runaway cost. A broken loop can burn tokens and tool calls until you hit a budget limit.

Module 5.4Experiment, then decide

Treat research as a pipeline. Try ideas safely, evaluate honestly, then adopt deliberately.

Open

Prerequisites

Comfort with earlier modules in this track
Ability to explain trade-offs and risks without jargon

Outcomes

Explain experiment, then decide in your own words and apply it to a realistic scenario.
Treat research as a pipeline. Try ideas safely, evaluate honestly, then adopt deliberately.
Check the assumption "Experiments are safe" and explain what changes if it is false.
Check the assumption "Evaluation is comparable" and explain what changes if it is false.

Practice

Solve a complex scenario with explicit assumptions and constraints
Write one mitigation plan and one fallback plan

Artefact and failure modes

A concise design or governance brief that can be reviewed by a team
Hype adoption. If you adopt because it is popular, you inherit failure modes you did not choose.
No exit strategy. If a new technique fails, you need a path back to the last safe version.

What changes at this level

Level expectations

Each level is independent but clearly deeper than the last. This panel makes the jump explicit.

Learning contract

Assessment outcomes

Read the explanation first, then use the tools to test the idea. Skip any tool that is not useful for your goal.

AI Agents stage assessment. Advanced mastery

Timed stage assessment with free access. Sign in so attempts and pass evidence stay attached to your account.

Stage 5 of 6

Practice before your timed attempt

Review your stage notes and rerun one practical tool before starting.
Complete the matching practice assessment for this stage.
Sign in and start the timed attempt when your reasoning is stable.

See sample questions Practice first Review stage content Sign in

Assessments record pass evidence and need your name. Sign in to continue.

Before you start

First-try attempts that reach the top 20 may be invited to publish to a child-safe public leaderboard.

You choose whether to publish. If you decline, your top 20 record is not saved.

AI Agents stage assessment. Advanced mastery

This assessment is timed. Question count and pass threshold are shown for your session when it loads.

Enter your full name. This name is locked for your pass records and CPD evidence.

You can retake as many times as you like. Attempts are tracked. Recent questions are avoided when possible.

What to Expect

Question Types

Multiple choice (single answer)
Code analysis questions
Scenario-based reasoning
Security vulnerability identification

Topics Covered

Fine-tuning and LoRA
Enterprise architectures
Production deployment
Research frontiers

Assessment Rules

Rules

The exact time limit is shown when you start the attempt.
Questions are drawn from the live published bank. The exact question count is shown when your session loads.
The pass threshold is 70 percent for this stage.
You can retake and attempts are tracked.
AI assistants are disabled during timed sessions.
Copy and context menu actions are restricted to maintain assessment integrity.

CPD candidates: Use the CPD prep pack for structured revision before attempting. If you need help, email ransford.amponsah@ransfordsnotes.com with your CPD reference code.

CPD prep pack Track CPD

Course materials are protected by intellectual property rights.View terms