Dimiter Todorov: I work across disciplines, and connect what they build.

Security · Integration · Automation · Architecture

I work across disciplines, and connect what they build.

Senior cyber security and integration specialist with 15+ years across datacenter automation, the Ontario government's first Azure hosting, red teaming and security operations, and now full-stack apps built with AI agents. I bring a generalist's view: architecture, integration, code and the screens people use, with security as the thread that runs through all of it.

Toronto, Ontario · Remote · Open to consulting and speaking

Dimiter Todorov: Shipped this year, with AI in the loop.

Top floor · Now building

Shipped this year, with AI in the loop.

A web app that finds the НЗОК entitlements a prosthetics center was missing, beside its legacy programs; an 8-day Remedy ↔ ServiceNow re-platform with 320 tests, an explainable risk model with its own workbench, a local ServiceNow emulator, and the handbook for doing AI-assisted engineering safely.

ServiceNow SDKTypeScriptKotlinReactClaude Code

Click a screen for the story behind each project.

On this floor

CheriBase: НЗОК rights for a prosthetics center

2026, ongoing · architect and lead developer

Allegro, a prosthetics and orthotics center in Bulgaria, ran on two legacy Windows programs and a shared Firebird database that could not say which patients were due a prosthesis, repair or consumables paid by НЗОК, the national health insurer. I led a web app that works next to those programs with byte-for-byte parity: an НЗОК rights engine with versioned rule sets the center's experts confirm, each entitlement explained in plain language, and a worklist that turns it into a call, a case and a claim. Microsoft Entra ID sign-in keeps every legacy change under a real legacy user. AI coding agents wrote most of the code, one per component, under contracts and build-enforced boundaries I reviewed. The first version is in use and has already brought real business results.

  • Kotlin
  • Ktor
  • React
  • TypeScript
  • Firebird
  • Entra ID
  • Azure
  • GitHub Actions
  • AI agents

Remedy ↔ ServiceNow integration re-platform

Oct – Nov 2026 · sole developer

A scoped ServiceNow app that brings Remedy incidents, CMDB, people and groups into ServiceNow through IRE, and sends Security Incident changes back. Built in 8 days with Claude Code: about 24K lines, 320 unit tests and 3 ATF suites, with the wiki written in the same commits. "Pure core, thin adapters": transform logic is platform-free TypeScript, tested against captured IRE contracts.

  • ServiceNow SDK / Fluent
  • TypeScript
  • IRE
  • BMC Remedy
  • Claude Code

Explainable vulnerability risk scoring

Nov 2026 · Vulnerability Response / USEM

Replaces raw scanner severity with a business-aware, seven-component risk model (severity, threat intel, asset criticality, data sensitivity, exposure, business impact, controls) where every score comes with a human-readable reason. A Risk Model Workbench lets you edit, validate, diff and preview a model on real CIs before activating it. New CI lookup rules substantially improved scanner-to-CMDB matching.

  • ServiceNow VR
  • USEM
  • Vue 3
  • Vitest
  • ATF

Klide: a local ServiceNow emulator

Sep – Oct 2026 · personal tooling

A Kotlin/Ktor service that runs ServiceNow server-side JavaScript on Rhino against PostgreSQL. It emulates GlideRecord, business rules, scheduled jobs, Scripted REST and IRE, verified against live responses, so apps can be built and tested offline with AI in the loop. 141 tests.

  • Kotlin
  • Ktor
  • Rhino
  • PostgreSQL
  • Vue 3

AI-assisted engineering, done safely

Nov 2026 · team enablement

The team handbook for agentic ServiceNow, Azure DevOps and Remedy work: AGENTS.md conventions, guardrails, reusable prompts and MCP usage, with one delivery path: prompt → human review → pull request → DEV install and ATF → evidence.

  • Claude Code
  • GitHub Copilot
  • MCP
  • Azure DevOps
  • GitHub Actions
Dimiter Todorov: Every finding on the right asset.

Integrations · the big one

Every finding on the right asset.

Microsoft Defender, Splunk and Sentinel, Recorded Future, Qualys, Guardium and BitSight into ServiceNow, with BMC Remedy as the enterprise CMDB. CI matching through IRE and lookup rules turns raw findings into context the risk model can use.

ServiceNow IREMID ServerMicrosoft GraphKQLBMC Remedy

Follow the packets: click the sources, the MID server and the hub.

On this floor

Remedy as the enterprise CMDB

Incidents, CIs, relationships, people, groups and the product catalog flow from BMC Remedy into ServiceNow through the Identification & Reconciliation Engine, with watermarks, paging and reconciliation jobs.

  • BMC Remedy
  • Atrium CMDB
  • IRE
  • MID Server

Microsoft Defender into ServiceNow

Defender TVM, XDR and Defender for Cloud findings land in Vulnerability Response and SIR. Pre-persistence filtering in the import pipeline projects about 40% less volume; a read-only Graph and KQL Advanced Hunting app enriches device context.

  • Defender TVM
  • Defender XDR
  • Microsoft Graph
  • KQL

SIEM into incident response

Splunk and Microsoft Sentinel ingestion into ServiceNow Security Incident Response, with Sentinel playbooks and phishing email-to-incident automation. Also Recorded Future threat intel, Qualys, IBM Guardium and BitSight.

  • Splunk ES
  • Microsoft Sentinel
  • Recorded Future
  • Qualys

The plumbing, as code

Azure-hosted MID Servers deployed with Bicep, managed identity and Key Vault, with certificate registration. Entra app registrations and GitHub OIDC federated credentials managed as code, so there are no long-lived secrets.

  • MID Server
  • Bicep
  • Entra ID
  • Key Vault
  • GitHub OIDC

One place for security operations

Ontario Treasury Board Secretariat · Cyber Security Division

Every finding lands on the right asset, with a risk score analysts can explain and an owner who can fix it. ServiceNow SIR, Vulnerability Response and USEM, connected to the tools the SOC already runs.

  • ServiceNow SIR
  • Vulnerability Response
  • USEM / CTEM
  • CMDB & CSDM

Context feeds the risk model

Asset criticality, data sensitivity, exposure and ownership come from the CMDB integrations, including database CIs modelled from Guardium, and become explainable risk scores refreshed daily.

  • Risk model
  • CMDB data quality
  • CI lookup rules
Dimiter Todorov: From red team to incident response.

Cyber security · threat intelligence

From red team to incident response.

Started in penetration testing and red-team tooling; then led the ServiceNow Security Incident Response and Vulnerability Response / USEM implementations for the Ontario Public Service Cyber Security Operations Centre.

SIRVR / USEM (CTEM)SentinelRecorded FutureMITRE ATT&CK

On this floor

Threat intel where analysts work

Recorded Future feeds into Splunk and CTEM enrichment, mapped to MITRE ATT&CK, so incidents and exposures arrive with context.

  • Recorded Future
  • Splunk
  • MITRE ATT&CK

Led Security Incident Response, end to end

Platform assessment (2023) → production

SIEM ingestion from Splunk and Sentinel, phishing automation, Remedy ticket synchronization, the SOC dashboard architecture, a critical-incident portal, and the SIR escalation integration with an external cyber partner.

  • ServiceNow SIR
  • Sentinel playbooks
  • Phishing automation

Started on the red team

2020 →

Red-team exercises, penetration tests and web/network vulnerability assessments across province-wide I&IT, with custom tooling, C2 and Splunk apps, and reports for senior management.

  • Penetration testing
  • Red team
  • Custom tooling

Microsoft 365 and Cloudflare

Microsoft 365 security and compliance configuration, and a full Cloudflare tenant set up and integrated with Entra ID for a medium-sized business operating in an EU jurisdiction.

  • Microsoft 365
  • Cloudflare
  • Entra ID
Dimiter Todorov: Agentic, with evidence.

AI

Agentic, with evidence.

Heavy, disciplined use of Claude Code and GitHub Copilot agents: guardrails in the repo, human review on every change, tests as proof. I wrote the team's AI-assisted engineering handbook, and built this site's engine the same way.

Claude CodeGitHub CopilotMCPSecurity Copilot

On this floor

Agentic engineering with evidence

Heavy, disciplined use of Claude Code and GitHub Copilot agents: guardrails in the repo, human review on every change, and tests as proof. That is how an 8-day integration re-platform shipped with 320 tests.

  • Claude Code
  • GitHub Copilot
  • MCP
  • AGENTS.md

Prompt → review → evidence

The delivery path from the AI engineering handbook: a prompt becomes a reviewed pull request, then a DEV install with ATF, and the evidence is kept. Also: Microsoft Security Copilot for phishing triage.

  • Azure DevOps
  • GitHub Actions
  • Security Copilot

Popfold, built with AI

This site runs on Popfold, a scroll-driven 3D engine built with Claude Code: the tower, the pipes, the studio and the Blender pipeline behind it.

  • three.js
  • TypeScript
  • Claude Code
  • Blender
Dimiter Todorov: Architecture people can use.

Design

Architecture people can use.

Solution and integration architecture, risk-model design, and Vue consoles for analysts. Popfold, the 3D engine running this tower, is portable: the same code powers a decorator's showcase and a consulting site.

Solution architectureVue 3three.jsPopfold

On this floor

Architecture first

Solution design documents, integration architecture, risk-model design and stakeholder workshops with ServiceNow, Splunk, Microsoft and IBM.

  • Solution architecture
  • Integration design
  • Workshops

Interfaces for analysts

Vue 3 consoles that make complex systems usable: an 11-view integration admin console, the Risk Model Workbench, and per-CI risk context panels.

  • Vue 3
  • UX
  • ServiceNow UI pages

Popfold: one engine, many sites

The 3D engine behind this tower is portable: the same code runs a decorator’s showcase with a folding compact, a consulting site in a modelled room, and this building, each with its own look, story and studio.

  • three.js
  • Design systems
  • Portability
Dimiter Todorov: 15+ years of integration and automation.

Ground floor · foundations

15+ years of integration and automation.

Datacenter automation across 8,000+ servers, the Ontario government's first Azure hosting, then security operations. Azure and Azure DevOps, Microsoft 365 security and compliance, and Cloudflare with Entra ID.

AzureAzure DevOpsMicrosoft 365CloudflarePowerShell

On this floor

15+ years of integration and automation

Ontario Public Service · 2009 – now

From datacenter operations (HP Server Automation across 8,000+ servers, runbook automation, closed-loop incidents) to the government’s first Azure hosting (GoCloud), to ServiceNow security operations. One thread: integration and automation around ITSM and the CMDB. LinkedIn

  • Treasury Board Secretariat
  • Ministry of Government Services
  • Georgian College

The toolbox

ServiceNow (SecOps and platform, SDK/Fluent, ATF, IRE), Microsoft Defender, Sentinel and Splunk, Azure (Bicep, Entra ID, Key Vault, Graph), Azure DevOps and GitHub Actions, Microsoft 365, Cloudflare. Languages: TypeScript, JavaScript, Kotlin, Python, PowerShell, KQL, SQL, Ruby, ClojureScript.

  • ServiceNow
  • Azure
  • Defender
  • Splunk
  • TypeScript
  • Kotlin

The CV

Every role, project and skill on two pages, as a PDF. It is made from the same data as this tower (src/cv.ts), so the two never disagree. Download the CV (PDF) LinkedIn

  • Open to consulting
  • Speaking
Dimiter Todorov: How it fits together.

The whole building

How it fits together.

  1. 1

    Security

    Know how systems break: red teaming, incident response, threat intelligence.

  2. 2

    Integration

    Connect the tools teams already run, so data arrives with context.

  3. 3

    Automation

    Pipelines, tests and AI agents do the repetitive work, with evidence.

Dimiter Todorov: Let's connect.

Let's connect.

Open to consulting and speaking on security integrations, ServiceNow SecOps and AI-assisted engineering. Toronto, Ontario · remote.