Real project work
You contribute to actual, in-progress projects — real data, real constraints — reviewed like production work, not marked like homework.
Training that behaves like the job
CapabilityNext exists to close the gap between knowing about AI and being trusted to build it. We do that with small cohorts, real projects, and mentors who work in the field.
Our approach
Build first, generalise later
Most people learn AI backwards — months of theory, then a scramble to apply it. We flip that: from week one you build something that runs, and the concepts arrive exactly when a project needs them. It all runs on GPU infrastructure we own, one click from your browser — so your attention stays on the learning, not the setup.
What makes a CapabilityNext cohort different
Mentor code reviews
Every submission gets written feedback from a practitioner, plus live review sessions.
Small groups
Roughly one mentor per eight students, so questions get answered the same week.
Our GPUs, one click
A browser workspace with datasets, PyTorch, Hugging Face and a real GPU pre-loaded — no installs, works from any device.
Responsible by default
Evaluation, bias, safety and cost are graded parts of every build — not an afterthought.
Portfolio to payroll
You leave with demonstrable work that maps to job requirements, a mentor reference, and a live role to apply for.
Who it's for
Students
Final-year and postgraduate students in CS, engineering, maths or another quantitative field.
Working engineers
Software engineers moving into machine-learning, data or AI roles.
Analysts & specialists
Domain experts who want to build AI systems, not just brief someone else to.
You should be comfortable with basic programming. Everything else, the two-week foundations sprint covers before your track begins.
2 wkfoundations sprint before every track
6–10 wkper specialisation course
3–5portfolio projects per graduate
1capstone you present and keep
Mentors
Who you'll learn from
Our mentors are working engineers and researchers. They rotate in from industry so the material tracks what teams actually do.
AK
A. Krishnan
ML Engineer · Recommender systems & rankingRM
R. Mehta
LLM Engineer · RAG, agents & evaluationSN
S. Nair
Data Platform Lead · Pipelines & MLOpsPT
P. Thomas
Applied Researcher · Computer vision
FAQ
Common questions
Do I need a machine-learning background?
No. You need to be comfortable writing basic code in some language. The two-week foundations sprint brings everyone up to speed on Python, the maths that matters and the tooling before your track starts.
How much time should I budget each week?
Plan for around 8–12 hours: live sessions plus project work. Cohorts run in the evenings and on weekends, and sessions are recorded so you can catch up.
Do I have to attend live, or online?
You choose. We run four delivery formats: fully self-paced (recorded), live online classes, a hybrid with live weekend sessions, and live in-person classes at our centre. The one-click GPU lab, the projects and the capstone are the same in all four — what differs is the amount of live contact, the schedule, the setting and the price. You can usually switch or upgrade format between modules.
Do you help with jobs?
Each programme is built around a live hiring need, so on completion there's a real role your portfolio maps to and you can start applying immediately. We also run interview preparation tuned to AI roles, review your portfolio and CV, and make introductions. We can't guarantee any individual offer — that still depends on your interviews.
Come build with us
Applications for the next cohort are open. A short form and a call is all it takes to get started.
Apply now