Software Engineering · AI & Agent Systems · Cloud/DevOps/Security · Research & Evidence Technology
This is the detailed portfolio behind the GitHub profile.
The GitHub profile is deliberately compact and visual. The portfolio is where the deeper material belongs: project units, technical demonstrations, architecture, implementation evidence, interactive experiments and development history.
Live site: https://abla86.github.io/developer-portfolio/
GitHub profile: https://github.com/abla86
Interactive entry points:
projects.html— searchable/filterable project atlas with project inspection panelsmission-control.html— Home / Engineering / Security mission-control interfacestack.html— interactive technology/competency matrix mapped to canonical repositoriesengineering-map.html— capability-to-evidence mapevidence-lab.html— interactive research/evidence workflow demonstrationssecurity-lab.html— bounded security demonstrationsdata-ocean.html— interactive data visualizationnetwork-simulator.html— interactive network simulationparticle-lab.html— interactive generative-computing experimentmatrix-wolf.html— interactive canvas experimentyggdrasil.html— interactive systems/visual experimentgame-lab/— modular interactive engineering engines
The intent is that every public page has a meaningful interaction, while the interaction remains technically bounded and does not replace the underlying source repositories.
The following Google Cloud completion badges are part of the evidence behind the AgentTrace AI-agent/security work. They are completion badges, not Google Cloud professional certifications:
- Introduction to Security Principles in Cloud Computing — completed September 22, 2026
- Strategies for Cloud Security Risk Management — completed September 22, 2026
- Secure Enterprise AI Agents — completion badge
- AI Boost Bites: Your Personal Feedback Agent — completion badge
- Build Collaborative Multi-Agent Systems with ADK & MCP — completion badge
- Optimize Agent Behavior — completion badge
- Agentic Strategy: Discover, Design, and Prototype — completion badge
- Secure Enterprise AI Agents — completed September 11, 2026
- AI Boost Bites: Your Personal Feedback Agent — completed September 6, 2026
- Build Collaborative Multi-Agent Systems with ADK & MCP — completed September 6, 2026
These badges are treated as training evidence. The corresponding applied evidence is the AgentTrace Google ADK/MCP integration, including the MCP policy gateway, tool-capability checks, provenance controls and audit trail.
- C# / .NET / ASP.NET Core
- Node.js / TypeScript / JavaScript
- React
- Python / FastAPI
- Java / Kotlin / Go
- REST APIs and backend services
- SQL / PostgreSQL
- automated testing and validation
- modular architecture
- Docker and local reproducibility
- event-driven architecture
- Apache Kafka
- service boundaries
- .NET Aspire
- Wolverine messaging
- polyglot consumers
- application integration
- agent orchestration and tool boundaries
- provenance and auditability
- policy enforcement
- prompt/input security
- source-grounded research workflows
- vector retrieval
- deterministic fallbacks and verification
- Microsoft Azure
- Kubernetes / AKS
- Docker
- GitHub Actions CI/CD
- Infrastructure as Code
- OIDC / Workload Identity
- observability
- DevSecOps
- dependency and static security analysis
- authentication and authorization
- document extraction and source provenance
- evidence appraisal workflows
- structured methodological checklists
- research traceability
- reproducible analysis
- explicit separation of software correctness from scientific validity
Healthcare is domain knowledge that informs selected projects; it is not the primary technology identity of this portfolio.
Projects are promoted according to public inspectability and evidence, not simply because they exist or use many technologies.
A public polyglot event-driven platform using a Node.js/TypeScript API, PostgreSQL, Kafka, Python, C#/.NET, Java, Kotlin and Go consumers, Docker, Kubernetes, .NET Aspire and Wolverine. This is the main integration point for the cross-language/full-stack requirements identified in target engineering roles.
Repository: https://github.com/abla86/AB-Engineering-Lab/tree/main/projects/eventforge
A reusable AI-agent security and provenance core with policy evaluation, tool-manifest integrity, audit evidence and deterministic testing. The flagship implementation now includes a specialist multi-agent PR review pipeline: Security, Dependency, Code Quality, Test and Risk/Feedback agents run behind an AgentTrace policy gate, with an optional Google ADK + MCP interoperability boundary. Specialist output is treated as untrusted tool output and cannot authorize a privileged write on its own. The implementation is deliberately synthetic and bounded; it demonstrates the architecture without claiming production deployment.
Google Cloud training applied here: cloud security principles, cloud security risk management, secure enterprise AI agents, agent feedback workflows, and collaborative multi-agent systems with ADK/MCP.
Repository: https://github.com/abla86/agenttrace
A cloud-engineering showcase covering .NET, React, containers, Kubernetes, Azure, IaC, CI/CD, DevSecOps, requirements, operational analysis and observability.
Repository: https://github.com/abla86/azure-kubernetes-showcase
A security-oriented portfolio verification system combining a React interface with a server-side verification and audit layer.
Repository: https://github.com/abla86/CodeSentinel
A healthcare technology API demonstrating C#/.NET, ASP.NET Core, backend services, database persistence, DICOM-oriented work, automated testing, Docker, CI and security boundaries.
Repository: https://github.com/abla86/HealthTechDeviceApi
A full-stack research-support application for structured appraisal workflows, methodology-aware instruments, traceability and review tooling.
Repository: https://github.com/abla86/complete-evidence-appraisal-tool
A full-stack workforce decision-support prototype using React, ASP.NET Core/.NET 10, SQL Server, Docker and automated verification. It models competence, availability, shift requirements and staffing coverage without presenting automated ranking as a substitute for professional or organisational responsibility.
Repository: https://github.com/abla86/workforce-competence-management
An AI-assisted publishing platform covering persistent book projects, Story Bible context, generation workflows, continuity auditing, versioning, creative assets, authentication, cost controls and auditability.
Repository: https://github.com/abla86/book-forge
A full-stack multi-agent software-engineering prototype with coordinated specialist agents, runtime telemetry, regression testing and bounded OWASP-oriented analysis.
Repository: https://github.com/abla86/Autonomous-Multi-Agent-App-Builder
A reusable research-engineering core for document extraction, source search, evidence provenance and integration with evidence-appraisal applications.
Selected supporting work demonstrates specific technologies such as data processing, retrieval/vector workflows, automation, interactive UI, finance/risk experimentation or cloud patterns. Small tutorials and historical experiments are kept out of the flagship set unless they provide distinct technical evidence.
The portfolio distinguishes:
prototype → demonstration → verified → production-ready
A repository is not described as production-ready merely because it builds or looks complete.
Where appropriate, the portfolio uses:
- source inspection
- automated tests
- CI
- linting and static analysis
- dependency/security controls
- documentation
- explicit architecture boundaries
- reproducible demonstrations
- live deployment verification where deployment exists
For research-support software, software correctness is kept separate from scientific or methodological validity. The application can automate a workflow, but scientific interpretation remains the responsibility of qualified users.
A technology from a job description is integrated into the repository where it has a legitimate architectural purpose. If integration would be artificial, it is kept as a bounded demonstration or explicitly marked as a future target.
The current mapping is maintained in AB-Engineering-Lab/docs/SKILL-INTEGRATION-MATRIX-2026-09.md and surfaced interactively through stack.html.
The portfolio is not intended to be a collection of static project cards. Public pages expose bounded interaction: filtering, search, inspection panels, simulations, visualizations, games and research/security workflow controls.
The interactive layer demonstrates implementation concepts; the GitHub repositories remain the authoritative implementation evidence.
Historical, training, archived and supporting repositories may remain visible on GitHub without being promoted as flagship products.
That distinction is intentional.
A project earns a primary position when there is enough:
- implemented functionality
- public inspectability
- verification evidence
- documentation
- distinct technical value
to justify the claim.
Build → test → secure → document → validate.
The objective is a portfolio that can be inspected by a technical reviewer rather than a collection of technology keywords.
The GitHub profile is the map.
The portfolio is the deeper system.
Repositories are the implementation evidence.
The visual tree on the profile represents capabilities and their relationship to the underlying work; it is not intended to replace the technical documentation.
Målrettet portefølje for rollen Systemansvarleg – IKT og helsedata: Åpne Helse Fonna-showcase
Denne siden samler de mest relevante tekniske bevisene: HealthTechDeviceApi, Complete Evidence Appraisal Tool, Azure Kubernetes Showcase og AgentTrace. Fokus er koblingen mellom helsefag, kunnskapsbasert praksis, systemforståelse, helsedata, sikkerhet og praktisk programvareutvikling.
Material portfolio changes should remain traceable to the relevant repository state and verification evidence.