3–5 weeks · one subject
Short, focused courses for working engineers. Take the one you need, not the eight months around it. Together the seven are the three-month bootcamp.
PML ──▶ DLP ──┬──▶ CV
└──▶ NLP
RAG ──▶ AGT no deep learning needed — start here if you build products
OPS needs a trained model of your own, from anywhereThe gap between a notebook model and one that ships.
Engineers who can train a model but have never had one audited, deployed or challenged in review.
BuildA model that would survive review — leakage-audited, cost-thresholded, SHAP-explained.
Train networks. Do not just call them.
Engineers moving beyond scikit-learn who need to build and debug their own training loops.
BuildA trained network with a full experiment log — curves, baseline, one ablation.
Detection, tracking and video, on real cameras.
Engineers building detection, tracking or video analytics on real footage and edge hardware.
BuildA vision system on live video — custom detector, measured mAP and FPS.
Fine-tune models. Do not only prompt them.
Engineers who need a model tuned to their own domain, language or task.
BuildA fine-tuned language model with a baseline comparison and error analysis.
Retrieval that actually retrieves the right thing.
Engineers building question-answering or knowledge systems over documents a business owns.
BuildA RAG system over a real corpus, with citations and a measured retrieval score.
Systems that decide and act, with guardrails.
Engineers putting tool-using LLM features into products people rely on.
BuildAn agentic application with tool calling, evaluation and a measured cost per query.
Ship it, run it, and know the moment it breaks.
Engineers responsible for models in production, and for the pager when they drift.
BuildA deployed, monitored service with CI/CD, model registry and alerting that fires.
All seven sprints in one document, with prerequisites and what each one ships. Download the PDF (16 pages).
Applications take ten minutes. We reply within three working days, either way — or just message us on WhatsApp.