Engagement: two enterprise-wide AI capability programmes (client and identifying details withheld under confidentiality). Shape: level-based tracks with placement diagnostics, capstones, mentoring, code review, and a monitored live period. Audience: technical practitioners from entry (L1) through principal/expert (L5), plus context authors, governance leads, and senior leaders in separate non-coding cohorts. Delivery: approximately 80% hands-on, in two-hour practical studios over multiple weeks to months.
Programme 1 · IT Team AI Enablement
Four progressive proficiency tracks — Builder to Principal
A · Builder (L1→L2)
~40–48 hrs over 10–12 weeks. Prompting and context engineering, agent architecture and function calling, Retrieval-Augmented Generation (RAG), automated evaluation and validation, AI safety/privacy and basic guardrails, inference/latency/cost measurement, LLMOps and foundational MLOps, and AI-assisted software development with coding assistants.
B · Specialist (L2→L3)
~48–56 hrs over 12 weeks. Dynamic context assembly and advanced prompting, orchestrator-worker and multi-agent design, agentic RAG across multiple sources, MCP server and tool design, human approval points and sandboxing, LLM-as-judge evaluation and CI release gates, role-based access/audit logging/advanced guardrails, and serving, tracing, fine-tuning foundations and ML-specific CI/CD.
C · Architect (L3→L4)
~56–64 hrs over 14–16 weeks, incl. capstone. Large-scale context architecture and automated prompt optimisation, peer and hierarchical multi-agent systems with fault tolerance, enterprise retrieval and semantic layers with lineage/freshness controls, MCP and agent-to-agent infrastructure, evaluation-driven development, compliance-aware guardrails/red-teaming/governance, production operations (blue-green, autoscaling, token accounting), LoRA/QLoRA or full fine-tuning and distributed training, and a defended capstone.
D · Principal (L4→L5)
~25–30 coaching hrs across 6 months, plus an independent charter. An individual research or innovation charter, advanced architecture, standards and technical leadership, research/conference/open-source contribution, mentoring and capability development for other practitioners, and long-term coaching against agreed milestones.
Programme 2 · AI Navigation Programme
A global intermediate-to-advanced programme — tool building through governance and leadership
1 · AI Tool Practitioner
~12–16 hrs over 3–4 weeks (entry via placement diagnostic; test-out credit for experienced users). Advanced prompt patterns and team prompt libraries, building assistants and applications with Copilot Studio or Dify, grounding/permissions/citation discipline, verification and responsible use of enterprise information, and applying AI to a real work process.
2 · Context & Instruction Design
4–6 week co-development sprint, plus a 10–14 hr Context Author certification. AI instruction templates and authoring standards; defining purpose, users, refusals and access boundaries; context curation, source ownership, refresh cadence and permissions; grounding rules, citations, confidence and escalation behaviour; and evaluation hooks, quality rubrics, prompt libraries and artefact registries.
3 · Agent Engineering
~25–40 hrs over 6–8 weeks, then a 30-day monitored live period. Agent design, delegation contracts, guardrails and exception paths; orchestration, tool integration, state and memory; sandboxed execution and human approval points; golden evaluation sets, rubrics and LLM-as-judge; CI/CD release gates; prompt-injection, data-exfiltration and tool-abuse testing; observability, tracing, cost, latency and reliability; and a defended capstone.
4 · AI Governance
~20–25 hrs. Multi-agent supervision, monitoring, audit trails and intervention; portfolio-level evaluations and review cadence; build/buy/wait decision-making; AI operating standards, incident response and rollback; AI performance scorecards; and a practical governance incident simulation.
Leadership Track (non-coding)
~8–12 hrs, separate senior-leader cohorts. AI policy and governance posture, AI programme strategy, build/buy/wait economics, reading AI performance scorecards and registries, and making evidence-based AI investment and risk decisions.
Why It Was Structured This Way
One-off awareness sessions don't build capability. These programmes are deliberately leveled and progressive — placement diagnostics meet people where they are, each track has a clear proficiency exit, capstones and monitored live periods prove real ability, and a separate leadership track equips decision-makers to govern and invest without writing code. The result is an organisation that can build, evaluate, operate, and govern AI on its own.
Related
← All Past Trainings
The full anonymized track record of delivered corporate AI trainings.
Corporate AI Training →
Team upskilling for engineering and L&D — build real in-house AI capability.
Agentic AI Training →
End-to-end agent stack with LangGraph, tools, memory, guardrails, and evals.
AI Adoption for Business →
Readiness, strategy, and governance for leaders deciding where AI fits.