Work I can show you.

An interactive Gatsby-style infographic and comprehensive reference guide for enterprise teams evaluating and adopting Claude across the 2026 model lineup. Covers the full stack: Claude Haiku 5.5, Sonnet 5.5, Opus 5.5 (released September 22 — the flagship for complex agentic coding and enterprise work), and Fable 5.1 (released September 1 — frontier intelligence for days-long autonomous work).

The portal includes model selection matrices, the latest releases (Claude Fable 5.1 on September 1 and Claude Opus 5.5 on September 22), the expanded four-exam Claude Certification Program (Associate, Developer, and Architect Foundations, plus Architect Professional), our Claude Partner Network membership, integration patterns (API, Claude Code, MCP), export control updates, Claude Tag announcement, and the PMP delivery touchpoints I rely on when scoping enterprise engagements.

Current through September 2026 — covers the Fable 5.1 and Opus 5.5 releases, Sonnet 5.5, the four-exam Claude Certification Program, US export controls, Claude Tag announcement, and enterprise certification programs.

• Demonstrates fluency with Anthropic's native artifact + Claude.ai workflow and current model ecosystem, backed by our Claude Partner Network membership.
• Shows how I translate vendor capability into client-ready reference material for fast evaluation.
• Bridges PMP discipline (scope, stakeholders, deliverables) with frontier-model tooling and enterprise deployment.
• Current through September 2026 with the latest model releases, export controls, and enterprise updates.

A daily situation report agent that researches, writes, and publishes a multi-domain morning brief — top AI story, What's Up in AI, TLDR AI, certification news, weather and tides — straight to theeglengroup.com.

The problem: the first version was a four-part AWS relay — EventBridge → Lambda (content + Polly audio → S3) → GitHub Actions → commit. When the Lambda ran late, the Action published a half-updated page: today's weather grafted onto yesterday's news.

What I built: in August 2026 I retired the relay and rebuilt the pipeline as a single Claude Code skill fired by a scheduled Routine — one agent, one author, from research through commit. Every section is sourced, missing sections are skipped rather than invented, and a personal copy is drafted for review, never auto-sent.

The result: 46 of 46 days published on schedule since cutover, with three moving parts (and a Lambda, an S3 bucket, and a CI workflow) removed.

• End-to-end agentic workflow: research, synthesis, publishing, and version control — in production, every day.
• A real migration story: diagnosing a race condition and replacing cloud glue code with a simpler agent-native design.
• Pattern is portable to any client needing executive situational awareness.

The agents that run The Eglen Group's own operations. Each one is a Claude Code skill — plain-English instructions with hard guardrails — scheduled as a Routine or fired on demand with a one-word alias. It is the same toolkit I bring to client work.

Scans job boards through the Indeed MCP connector and web search for enterprise postings that name Claude or Anthropic, reads each full description to confirm the match (and to rule out "Claude" the hiring manager), dedupes against a state file in git, and drafts a Gmail digest of what's new. Along the way it found that Indeed's job IDs change on every request, so it keys on a normalized company · title · location instead. The same pattern works for lead generation, RFP monitoring, and competitor watch.

Replaced a GitHub Actions job that called the Anthropic API from CI with a native Claude Code skill that refreshes stale dates, stats, and seasonal references across eleven pages. Its hard rules forbid touching HTML structure, classes, styles, or scripts, and forbid inventing facts: if it isn't sure something is stale, it leaves it alone.

One source of truth per pipeline, thin aliases that call it, known gaps written into each skill, and human review before anything leaves the building. Drafts, not sends.

• Working agents in daily use, not slideware, and the git history to prove it.
• Shows the guardrail thinking enterprises need before they let an agent near production.
• Each pattern maps directly to a client engagement: monitoring, content operations, reporting.

A working reference for Agile delivery on AI/ML programs — grounded in the Agile Alliance Manifesto, PMI-ACP frameworks, and adapted for the non-deterministic realities of generative and agentic AI.

Covers the 4 values and 12 principles, the March 2026 PMI-ACP Exam Content Outline, adapted sprint ceremonies for GenAI, the Minimal Viable Agent pattern for agentic delivery, framework selection guidance, and a full toolkit of AI project templates.

• Demonstrates that delivery methodology is as central to the practice as model selection.
• Shows how PMI-ACP discipline and Agile Alliance principles are adapted — not abandoned — for AI programs.
• Addresses the most common failure mode in enterprise AI: strong model, undisciplined delivery.
• Directly applicable to client engagements — this is the methodology framework I bring to every program.