Inference Engineer roadmap
Six stages · about four months at six hours a week · 25 resources
Make machine learning models fast, cheap and reliable in production.
Leads to roles such as Inference Engineer, ML Systems Engineer.
Create a free account to open the lessons and videos, tick off stages, save this roadmap and track your progress.
The stages
Stage 1: Model and hardware basics
What happens when a model runs: memory, compute and latency.
about 3 weeks 5 lessons and videos inside
You will be able to
- Explain latency versus throughput
- Estimate a model's memory needs
- Read a simple profiler trace
Project
Measure the latency of a small open model on your machine.
Stage 2: Serving a model
Wrapping a model in an API and handling concurrent requests.
about 3 weeks 4 lessons and videos inside
You will be able to
More roadmaps to explore
- Data AnalystSQL, spreadsheets and telling the truth with numbers.
- AI & Machine LearningFrom Python and maths to models you can explain.
- Data ScienceStatistics, Python and machine learning to answer real questions.
- AI EngineerBuild useful products on top of language models.
- Data EngineerMove, clean and organise data so others can trust it.
- MLOps EngineerTake machine learning models from a notebook to a reliable service.