TT Claude Memory

project_api_days_singapore_contacts

Properties1
descriptionKey contacts from Rich's attendance at apidays Singapore (~April 2026). Donnie (AWS Developer Platform & AI Lead, ASEAN) introduced Rich to Erik Wilde (OpenAPI Initiative Ambassador, Jentic). Strategic implications for INHERIT's API layer and agent-readiness.

5 min read

Donnie Prakoso — AWS Developer Platform & AI Lead (Singapore)

  • Developer Engagement Unit Lead + Principal Developer Advocate at AWS, ASEAN region
  • 20+ years distributed systems + cloud; former CTO
  • Led GTM for AWS Region launches in Indonesia, Malaysia, Thailand
  • Designs GTM for Amazon Q Developer, Kiro IDE, Amazon Bedrock across 1.45M developers
  • Very enthusiastic about OpenAPI for agentic work — views OpenAPI as critical for AI agents
  • Building his own LLMs — easier to control, cheaper to run (worth noting as a pattern; not InheritKit-specific yet)
  • Created Spec-First TDD methodology (combines SDD + TDD with AI coding assistants)
  • Monthly AWS ASEAN Devs/Communities newsletter (5,446 subscribers)
  • LinkedIn: linkedin.com/in/donnieprakoso/

Context of the meeting: Donnie expressed interest in using INHERIT with an Indonesian freight client. Rich correctly told him there are better options for that problem — scope-integrity move that preserved INHERIT’s credibility. Donnie then introduced Rich to Erik Wilde.

Erik Wilde — OpenAPI Initiative Ambassador, Head of Enterprise Strategy at Jentic

  • OAI Ambassador at OpenAPI Initiative since 2023
  • Co-author of RFC 9457 (Problem Details for HTTP APIs)
  • Author of Continuous API Management (O’Reilly, 1st ed 2018, 2nd ed 2021)
  • Dr. Computer Science, ETH Zürich (1997)
  • 15+ years in API governance, design, strategy
  • Former API Catalyst at Axway; Principal Consultant at INNOQ
  • Runs “Getting APIs to Work” consultancy/YouTube channel
  • Since Sep 2025: Head of Enterprise Strategy at Jentic
  • LinkedIn: linkedin.com/in/erikwilde/
  • Zürich-based

Jentic — agent-ready API infrastructure

Founded late 2024/early 2025. Key team:

  • Dorothy Creaven — Co-Founder & COO, Board Director
  • Michael Cordner — CTO & Co-Founder
  • Erik Wilde — Head of Enterprise Strategy
  • Frank Kilcommins — Head of Enterprise Architecture

Product focus:

  • Published open-source AI-Readiness Framework for APIs (scoring model)
  • jentic/api-problem-details — reusable RFC 9457 Problem Details components for OpenAPI, Python, TypeScript (Apache 2.0)
  • Works on spectrum of LLM/agent control: MCP (expose tools) → Agent Skills (packaged tasks) → Workflows (deterministic multi-step, Arazzo-based)
  • Frank Kilcommins thesis: structured machine-readable errors required for agent recovery — aligns with RFC 9457

Strategic relevance to INHERIT:

  • Jentic’s AI-Readiness scoring model is direct external evidence for what agent-consumable APIs need — must feed into foundation-architecture Phase 1
  • OpenAPI + Arazzo is the API + workflow description layer INHERIT needs regardless of its underlying data format (JSON-LD, RDF, or JSON Schema)
  • Erik as OAI Ambassador is a potential institutional ally for INHERIT’s API layer — not a competitor
  • RFC 9457 Problem Details is a pattern INHERIT should adopt for error responses
  • Jentic’s spectrum (MCP/Agent Skills/Workflows) is directly relevant to how InheritKit exposes its operations to agents

Rich’s emotional context at meeting

Rich was nervous about meeting Erik because he felt his own OpenAPI expertise was limited. This was misplaced — Erik evaluates substance in one’s domain, not API terminology fluency. Rich’s domain (legal succession, 50+ jurisdictions, multi-faith) is genuinely novel; Erik’s domain (API specification and governance) is complementary not competitive. Future interactions: Rich approaches as domain expert seeking perspective on API layer, which is a peer-to-peer frame, not asymmetric.

Action items (parallel to foundation-architecture research)

  1. Reach out to Erik post-apidays with a thank-you and a “I’d value your perspective on INHERIT’s API layer” opener. Low-pressure, authentic.
  2. Follow Jentic’s open-source work — particularly the AI-Readiness Framework and Problem Details components. Both directly applicable to INHERIT infrastructure.
  3. Track Donnie’s work on LLMs, Spec-First TDD, and Kiro IDE. Monthly newsletter is cheap keep-in-touch.
  4. Consider Jentic as a potential technical partner for INHERIT’s API layer, if the foundation-architecture research confirms OpenAPI + Arazzo is part of the stack. Analogous to Sourcemeta+Blaze for JSON Schema, but for agent-ready APIs.

The v2.4 foundation-architecture handoff prompt adds OpenAPI / Arazzo / agent-readiness as a fourth evaluation dimension (joining internationalisation, standards-body recognition, and core expressivity). Donnie and Erik’s perspective is a direct signal that agent-consumable-ness is a 20-year-viability criterion INHERIT cannot skip. The Jentic AI-Readiness Framework is explicit Phase 1 subagent reading.

Flag closed — 2026-04-22 v1.9 architecture-proposal

The “OpenAPI + Arazzo + agent-readiness missing from foundation thesis” flag is now ADDRESSED. Rich raised it during the preparatory session for the 24-Z research programme; agent pushed back that OpenAPI had drifted to “v6.6 inherited reference material” rather than first-class architectural concern. Architecture-proposal v1.9 §“OpenAPI + Arazzo agent-contract layer” now promotes OpenAPI 3.1 + Arazzo 1.0 to the Ⅲ stack. Changes:

  1. Ⅲ stack includes OpenAPI 3.1 + Arazzo 1.0 alongside Catala + JSON Schema + JSON-LD + SPARQL + SHACL + OWL 2 DL + InheritKit
  2. Per-module OpenAPI specs — 7 published specs under code-inherit-v2/modules/{module}/openapi.yaml
  3. Arazzo 1.0 workflow library — 7 foundational workflows Phase-1 scope
  4. InheritKit MCP auto-generated from OpenAPI — follows Sourcemeta One precedent at Apache tier; TT-built at AGPL+commercial tier
  5. AI-vendor commercial tiers (Bronze/Silver/Gold) — makes project_ai_vendor_commercial_model operationally concrete; licensees consume OpenAPI + Arazzo + hosted MCP
  6. Jentic institutional-ally opportunity re-activated — Erik Wilde as OAI Ambassador + Jentic AI-Readiness Framework + RFC 9457 Problem Details pattern all feed INHERIT v2 agent-readiness

Phase-1 LOC envelope grew ~45-55K canonical → ~60-85K canonical to accommodate. Timeline 30-48 → 33-54 months. Slip-trigger adjusted: Arazzo workflow library is first-drop if velocity underperforms.

Z3-AGENT + Z9-STANDARDS + Z16-INTEGRATIONS handoff v1.5 addenda expanded with OpenAPI-explicit research scope. Library already covers via ponelat-designing-apis-openapi + itelman-unifying-business-data-code + w3c-mcp-spec + anthropic-mcp-roadmap-2026 + gulli-agentic-patterns. Optional acquisition: Arazzo 1.0 spec (free from OAI).

Action item #1 (reach out to Erik post-apidays) remains active. Under v1.9, the OpenAPI-first positioning strengthens the approach — Rich can now say “we are OpenAPI-first from v2.0” rather than “we’re considering OpenAPI.”

Graph View

Backlinks