Skip to main content
Skip to content
Menu

AI pretraining Intermediate 3 parts

Compose a brief internal memo recommending which pre-trained model to adopt for your team

You'll have a 1-page memo with model comparisons (performance, cost, latency, license), ranked by your team's top 3 priorities, with one clear recommendation and a deployment plan.

What you'll make

Model Selection Memo

  1. Executive Summary: one model recommendation with deployment timeline
  2. Comparison Matrix: 3-4 models scored against your team's top 3 priorities
  3. Testing Results: specific performance on your team's actual tasks
  4. Deployment Plan: concrete next steps and risk mitigation

Write as a memo you'd actually send to your manager or team lead

The 3 parts

Each part ends with a check you can see.

  1. Map Your Team's Real Constraints

    Identify the 3-4 factors that will actually determine model success for your team's work.

    Done when

    You have ranked constraints (cost, latency, accuracy, deployment complexity) with specific thresholds your team can't cross.

  2. Test Models Against Real Tasks

    Run 3-4 candidate models on your team's actual work and measure what matters.

    Done when

    You have test results comparing specific models on your prioritized criteria with concrete numbers and failure examples.

  3. Draft Your Recommendation

    Write the memo with your model choice, supporting data, and deployment plan.

    Done when

    You have a complete 1-page memo that makes a clear recommendation with evidence and addresses likely stakeholder concerns.

    Capstone: you put together your Model Selection Memo.

Where this helps

An elective: a single lesson that sits outside rtwo's job certificates and doesn't count toward one.

About this lesson

What you'll be able to do

  • Write a defensible model selection memo based on actual testing, not vendor claims
  • Compare 3-4 models on metrics that matter to your team's real work
  • Have a deployment plan with concrete next steps and risk mitigation

Who it's for

  • Technical leads deciding between model options for their team
  • Engineers who need to justify model choices to stakeholders
  • Anyone who's tired of choosing models based on leaderboard scores instead of real performance

What it won't cover

  • Fine-tuning or training models from scratch
  • Deep technical architecture comparisons
  • Multi-model orchestration or ensemble strategies
  • Production monitoring and observability setup

Skills you'll practise

  • Comparative model evaluation
  • Business constraint mapping
  • Technical risk assessment
  • Stakeholder-facing technical writing