Training
Programs for engineering teams, product leaders, and executives at regulated institutions, taught from the systems we operate. Every module traces back to something we run in production.
Programs
AI for the Enterprise
For teams putting language models into production: model selection, context engineering, integration, and the governance evidence a regulated deployment needs.
Taught from our deployment work with banks and enterprise software companies.
What’s covered
Models & Selection
How current models differ, what they cost, and when to choose API, fine-tuning, retrieval, or agents, judged by production tradeoffs rather than benchmarks.
Context Engineering & Retrieval
Retrieval, structured context, and memory patterns that hold as documents, tools, and users multiply.
Integration Architecture
API design, failure handling, fallbacks, and cost control for models under real traffic.
Evaluation & Governance
Evaluation pipelines your model risk team will accept, from the challenge-driven method behind Versalist.
Agentic AI & Multi-Agent Systems
For teams building agents that take real actions: architecture, tool use, orchestration, and the operating discipline an institution needs before agents touch production.
Taught from the agent infrastructure we operate.
What’s covered
Agent Architectures
Planning loops, reflection, multi-step execution, and when a simpler pipeline beats an agent.
Tools, Permissions & Authority
Tool interfaces and permission boundaries, what an agent may see, do, and commit: the authority model behind KnowYourAgent’s trust rails.
Orchestration & Command
Directing many agents as a small team: coordination, shared state, and failure isolation, from operating ActsAsGeek.
Production Operations
Monitoring, human-in-the-loop controls, and append-only audit trails: answering for an agent’s actions after the fact.
AI-Augmented Software Development
For engineering organizations adopting coding agents in a regulated SDLC: the productivity without losing the evidence trail examiners ask for.
Taught from AutoDevOps, the evidence pipeline built for this problem.
What’s covered
Working With Coding Agents
Session design, context management, and review workflows for Claude Code, Cursor, and Codex.
Policy & Permission Gates
What an agent may run, confirm, or never touch, enforced before the action executes.
Evidence & Audit
Append-only logs, evidence packages, and model risk thinking for AI-written code: evidence an examiner can verify.
Quality & Security Gates
Tests, review discipline, vulnerability patterns, and the failure modes that pass CI.
Format & Delivery
- Led by a practitioner who has built and shipped these systems at scale.
- Customized to your stack and regulatory context. Tailored to your technology environment, examination posture, and team maturity.
- Hands-on. Participants build working prototypes on their own tools and infrastructure.
- On-site or remote. Half-day, full-day, and multi-day formats, single or as a series.
Bring a program to your team
Tell us your team, stack, and regulatory context. We’ll design a program that fits.
Request a Program