Skip to main content

Applied AI Fellowship

A production-grade Applied AI capability framework designed for working professionals in enterprises and governments.

Why This Course Exists​

After two decades of leading product and engineering teams across industries - from climate tech to blockchain, from startups to large enterprises - I've seen too many AI projects fail. Not because the technology wasn't capable, but because teams lacked the foundational knowledge to make sound decisions, build safely, and deliver value.

In my current role leading AI transformation at a major university, I'm driving AI SDLC transformation, mentoring teams to democratize data, and spearheading AI applications across the organization. This experience has reinforced a critical gap: there's a chasm between AI hype and production reality.

Value proposition: This fellowship exists to bridge that gap. It's born from the need to equip professionals with practical, production-grade capabilities - not just theoretical knowledge or prompt engineering tricks, but the judgment, rigor, and governance skills needed to build AI systems that work, scale, and earn trust.

Whether you're an executive making investment decisions, a product leader shaping agentic roadmaps, or an engineer shipping production agents, this program provides the structured learning path to develop real capability. Each track is designed for a specific audience, with multiple levels that build from foundation to mastery.

That path is vendor-neutral by design. The AI Capabilities Stack, enterprise planes, and track curricula are framed around architecture, risk, and governance - what stays true whether you deploy on Microsoft, Google Cloud, or AWS, or with models and tooling from OpenAI, Anthropic (Claude), or others. Representative tags in the stack point to real products and protocols so you can map ideas to names you already know; they are not a single-vendor syllabus.

Vendor-specific training still matters. Cloud and model-provider courses teach you their consoles, APIs, reference architectures, and supported patterns so your teams can ship quickly on the stack you have committed to. This fellowship does not replace that. It sits alongside it: vendor curriculum answers "how do we use this platform well?"; this program answers "how do we decide, govern, and structure AI so we are not reinventing risk management every time the logo on the invoice changes?"

Organizations thrive when both layers exist. Principled approaches - clear ownership, control outside the agent loop, observability, cost discipline, memory versus retrieval, and honest enterprise trade-offs - keep you from mistaking a vendor roadmap for a strategy. Combined with Microsoft, Google, AWS, Claude, OpenAI, or other vendor depth on your teams, you get speed where you have contracted and judgment that survives platform churn. That pairing is why this course exists in addition to, not instead of, vendor-specific education.

What Makes This Different​

  • No hype - Focus on practical, production-ready AI
  • No prompt-only learning - Deep understanding of AI systems
  • Governance-first - Policy, identity, tool registry, budget enforcement, and progressive delivery live in the control plane - outside the agent loop, deterministic and not prompt-based (same framing as the AI Capabilities Stack)
  • Stack-grounded - Eight agentic stack planes (UX, Control, Runtime, Context & Memory, Data, Integrations, Security, Observability) as a risk surface map, plus five enterprise planes (people, culture, governance, strategy, economics) for sponsors and executives - see the AI Capabilities Stack
  • Same synthesis as the diagram - Pedagogy (Bloom, artifact-driven, role-specific, governance-first); technical depth (production patterns, build-vs-buy, 2026 tooling, plane architecture); memory ≠ RAG; governance outside the agent; integrations maturity as Integration / ESB → Data Engineering → RAG → Memory → (forward: AR/VR (emerging)); cross-cutting AI SDLC and cost management
  • Artifact-driven - Real deliverables, not just theory

Our Methodology​

This fellowship is built on proven educational frameworks - Bloom's Taxonomy, knowledge dimensions, artifact-based assessment, and explicit focus on transfer and judgment. Learn more about our methodology →

Program Overview​

The Applied AI Fellowship is organized into five role-specific tracks. You follow one track for your role; lengths reflect depth and scope: 4 weeks (T1 Executive, T2 AI Product), 8 weeks (T3 Delivery Leader), 10 weeks (T4 AI Engineer), and 12 weeks (T5 AI Architect). Each track contains multiple levels with defined learning outcomes so you progress from foundational to advanced capability within that role.

Five Learning Tracks​

Choose the track that matches your role and goals:

  1. T1: Executive - Control plane: governance, procurement, and investment - policy and budgets outside the agent loop
  2. T2: AI Product - UX (what users see, approve, and download; maturity borrowed surfaces → dedicated hub → hybrid) and control (pricing, progressive delivery)
  3. T3: Delivery Leader - UX (HITL / approval UX), control, runtime (contracts, SLAs, multi-lane literacy), observability (delivery lens) - planes as risk framework
  4. T4: AI Engineer - Control (MCP tool registry, guardrails, identity, budget), runtime (multi-lane engineering), integrations (tooling + MCP), security (guardrails, zero-trust), observability (OTel, eval, cost attribution)
  5. T5: AI Architect - All eight planes with depth on Context & Memory (four networks: World, Experience, Opinion, Observation; TEMPR (Temporal Entity Memory Priming Retrieval) via Hindsight; Astrocyte-augmented patterns), Data (multi-tenant ownership, behavior schemas, schema as source of truth), Integrations (protocol convergence MCP → ACP → A2A), and Security (incl. CA SB 243/AB 489 where relevant)

Program Structure​

  • One track per fellow - Durations above; not every role needs the same calendar length
  • Eight agentic stack planes - Same labels and intent as the diagram: UX (trust surfaces; approve / download); Control (policy, identity, tool registry, budget, progressive delivery); Runtime (multi-lane: frontdoor, durable worker, sandbox, GPU inference - each lane: latency, isolation, cost); Context & Memory (cognition vs RAG-only; four networks + TEMPR (Temporal Entity Memory Priming Retrieval) with Hindsight; vectors vs graph stores, hybrid); Data (multi-tenant ownership, behavior schemas, tool I/O contracts); Integrations (search, APIs, drift; MCP → ACP → A2A); Security (identity, I/O guardrails, zero-trust workload identity); Observability (traces, metrics, eval, cost; OTel GenAI semantic conventions; three eval strategies). Five enterprise planes (people & leadership, culture & change, governance & risk, strategy & portfolio, economics & ROI) sit beside them for org, culture, and financial framing. The AI Capabilities Stack is the interactive source of truth.
  • Footer bars (map) - Stack maturity: Integration / ESB → Data Engineering → RAG → Memory → AR/VR (emerging); Enabling: AI SDLC + Cost Mgmt · Enterprise planes: People & Leadership → Culture & Change → Governance & Risk → Strategy & Portfolio → Economics & ROI (second row; complementary to the technical chain, not a duplicate of it)
  • Memory vs retrieval - RAG optimizes retrieval; memory systems support ongoing cognition - an explicit distinction in the architecture track
  • Multiple levels per track - Progress through levels with defined learning outcomes
  • Artifact-based assessment - Build real, reusable deliverables
  • Credentialed badges - Earn industry-recognized certifications
  • Public sector variant - Specialized track for government and regulatory contexts

Outcomes​

  • Better decision-making - Know when to invest in AI and when to avoid it, reducing wasted effort on unsuitable projects
  • Higher success rate - Governance-first approach increases project success
  • Stronger trust - Built-in accountability and auditability

Getting Started​

Explore the learning tracks to find your path (each track page includes its Schedule section), see the AI Capabilities Stack (tracks × planes × curriculum), use the Glossary for canonical product and protocol names, learn about our methodology, or discover credentialing and badges.

Ready to get started? Get in touch to learn more or apply.