Advanced AI Specialisations
Explore five AI engineering tracks: advanced agents, multimodal, computer vision, voice, and language AI, then build one starter project in depth.
What You'll Learn
Choosing Your Specialisation Track
By now acme-assist answers questions from the help centre, issues refunds safely, and has a release gate.
Going Deeper on Agent Engineering
This track is for you if building the refund agent was the part of the roadmap you enjoyed.
Building Multimodal Document Extraction
Multimodal AI does not mean image generation.
Building Computer Vision Counting
Computer vision is the track for precise, structured, real-time visual understanding at a scale and cost general models are not built for.
Building Voice AI Pipelines
Voice AI is the track for spoken interfaces: phone support, in-app assistants, replacing IVR menus.
Building Language AI Systems
Language AI, often called NLP, is the track for teams handling huge volumes of text where an LLM call per item is too slow, too expensive, or...
Skills You'll Master
Curriculum Index8 topics
Choosing Your Specialisation Track
By now acme-assist answers questions from the help centre, issues refunds safely, and has a release gate.
Going Deeper on Agent Engineering
This track is for you if building the refund agent was the part of the roadmap you enjoyed.
Building Multimodal Document Extraction
Multimodal AI does not mean image generation.
Building Computer Vision Counting
Computer vision is the track for precise, structured, real-time visual understanding at a scale and cost general models...
Building Voice AI Pipelines
Voice AI is the track for spoken interfaces: phone support, in-app assistants, replacing IVR menus.
Building Language AI Systems
Language AI, often called NLP, is the track for teams handling huge volumes of text where an LLM call per item is too...
Understanding Reinforcement Learning at Awareness Level
Reinforcement learning is deliberately not a track this roadmap expects you to master.
Reviewing the Quick Reference and Common Mistakes
Quick reference Common mistakes Choosing a track because it sounds impressive rather than because of the company you...
Career Impact
Roles that use the skills in this module.
AI Engineer
MLOps Engineer
Platform Engineer
Next Modules
Practice on the Coding Sheet
Not a software engineer sheet. Every problem comes from real DevOps, SRE, Platform and Cloud interviews, from your first script to a system you build yourself.
Open the Coding SheetFrequently Asked Questions
No. Pick one, build its starter, and go deep. Interviewers can tell one working system with handled edge cases from five demos that break on the first unusual input. Keep working knowledge of the others.
Choose by the kind of product and company you want to work on, not by what sounds impressive. Text classification at scale points to Language AI, phone support points to Voice AI, and warehouse cameras point to Computer Vision.
No. Every starter in this module runs on a laptop CPU for free. The optional go-further steps, such as fine-tuning a detector on a dense-shelf dataset, are the parts that benefit from a GPU, and each says so.
No. It is covered at awareness level only. It matters for specialised ML and research roles, and an AI engineer mainly needs to know what RLHF is and why chat models behave differently from raw pre-trained ones.