Skip to content
Back to selected workmydecisive.ai · Jul 2026

Octant

Enterprise AI DevOps commercial suite

The problem

Leadership proposed a four-week delivery window for work involving a new UI, APIs, continuous-delivery integration, platform integration, and new engineering contracts.

What I owned

  • Helped conceptualize Octant from the ground up.
  • Authored all Octant documentation and most of its architecture designs.
  • Wrote product one-pagers that converted platform and customer problems into scoped technical initiatives.
  • Designed how Octant would be delivered through the existing platform, tooling, technology stack, and integration points.
  • Established telemetry validation as an architectural requirement for measuring fidelity loss when vendor-native agents and collectors translate through OpenTelemetry.
  • Implemented a JSON-driven frontend content layer so marketing and product collaborators could edit Octant content without routine code changes.
  • Deliberately stayed out of core implementation to preserve independence for a true customer-perspective end-to-end evaluation.
  • Manually exercised Octant end to end from a customer perspective and drove repeated bug-fix iterations before launch.
  • Decomposed the initiative and created engineering estimates.
  • Identified dependencies and prioritized deliverables.
  • Proposed phased delivery and negotiated scope from a four-month engineering estimate to an approximately three-month delivery plan.
System design

How the architecture works.

Architecture view

Architecture and independent validation as launch controls

Technical framing, delivery architecture, and integration requirements established the system boundaries; independent customer-perspective testing then drove stabilization before commercial launch.

Drag or use arrow keys to pan · pinch or Ctrl/⌘ + scroll to zoom

50%

Read the architecture as text
  1. Commercial initiative + platform constraints — decompose → One-pagers + scoped technical initiatives
  2. One-pagers + scoped technical initiatives → Delivery architecture + integration points
  3. Delivery architecture + integration points — define validation → Multivendor telemetry-fidelity requirements
  4. Delivery architecture + integration points — create collaboration boundary → JSON-driven frontend content layer
  5. Multivendor telemetry-fidelity requirements → Independent customer-perspective E2E testing
  6. JSON-driven frontend content layer → Independent customer-perspective E2E testing
  7. Independent customer-perspective E2E testing → Bug-fix iteration + tagged releases
  8. Bug-fix iteration + tagged releases — verify fixes → Independent customer-perspective E2E testing
  9. Bug-fix iteration + tagged releases → Commercial launch
Explore the design decisions and capabilities

    implementation scope

    Jillian intentionally limited her production-code contribution to the frontend, where she created easy-to-edit, CMS-like JSON files for marketing and product collaborators. She did not implement the core Octant system.

    Separating architecture and evaluation from core implementation allowed her to remain objective and exercise the complete product as a customer would during end-to-end usage testing.

    customer perspective e2e

    Manually tested the complete Octant experience as a customer would. Jillian reports that nearly all tagged pre-launch iterations were connected to bugs she exercised and verified through this process.

    delivery architecture

    Designed the architecture connecting Octant to the existing MDAI platform, its delivery tooling and technology stack, and the required system integration points.

    product architecture

    Authored one-pagers and architecture proposals covering configurable Octant experiences, telemetry-attribute discovery, operational health and insights, and reusable foundations for guided and agent-driven workflows.

    multivendor telemetry fidelity

    Identified that translating vendor-native telemetry through OpenTelemetry is a semantic round trip rather than a transparent proxy. Required explicit semantic ownership, sampling authority, documented fidelity limits, and automated validation across metrics, logs, and traces before an integration could be considered stable.

    Service-name stability
    Environment and version propagation
    Host attribution
    Error tagging
    Sampling correctness
    Log-trace correlation
    Metric-tag integrity
    Evidence

    What changed.

    • Enabled marketing and product collaborators to modify frontend content through a JSON-driven content layer without depending on routine engineering changes.
    • Preserved an independent customer perspective for end-to-end product evaluation by intentionally separating architecture and testing responsibilities from core implementation.
    • Drove a dense pre-launch stabilization cycle through manual customer-perspective end-to-end testing.
    • Reduced launch risk and support burden; no customer-reported issues or public GitHub issues were recorded after release as of 2026-07-26.
    • The negotiated approximately three-month plan shipped as Octant.
    • Octant reached public commercial launch and general availability in July 2026.
    • The launch received independent coverage in SiliconANGLE alongside syndicated coverage of MyDecisive's press release.