Technical Product, Platform & AI Services

I work across technical product design, platform architecture, software engineering, AI integration and structured delivery. The value is not in treating these as separate services, but in connecting them from requirements and system boundaries through implementation, deployment and operation.

Core Service Areas

Technical Product & Solution Architecture

Translate product goals and requirements into an implementable technical design. This can include system boundaries, non-functional requirements, data models, API contracts, integration strategy, identity and authorization, AI boundaries, deployment topology and architecture decisions.

Platform Architecture

Design reusable technical foundations rather than solving the same problem independently in every feature. Relevant areas include multi-tenant systems, shared services, permission models, structured publishing, provider abstraction, modular APIs, multilingual architecture and reusable AI capabilities.

Software Engineering & Full-Stack Implementation

Implement web applications and platform capabilities across frontend, backend, databases and APIs. Current work commonly uses Nuxt, Vue, TypeScript, Node.js, Nitro, Prisma and PostgreSQL, with Python and other technologies where they are the better technical fit.

• Business and internal applications
• Customer and partner portals
• Administration interfaces and dashboards
• Multi-tenant and SaaS-oriented systems
• APIs and integration services
• Workflow and document applications
• Structured publishing platforms
• Commerce and product-oriented platform components

AI Integration & AI Systems Architecture

Integrate AI as part of an application architecture rather than as an isolated API call. This includes hosted and local models, provider abstraction, retrieval, tools, agents, permissions, structured outputs, evaluation, observability and controlled runtime behavior.

• Local, hosted and hybrid AI architectures
• Model and provider abstraction
• RAG and external knowledge integration
• Agent and tool-based workflows
• Local AI clients and controlled runtimes
• Structured extraction and classification
• AI-assisted publishing and knowledge workflows
• Application-level validation and failure handling

RAG, Search & Knowledge Systems

Design retrieval and knowledge layers that connect applications and AI systems with external information. This can include ingestion, chunking, metadata, lexical and semantic search, embeddings, reranking, provenance, Source-of-Truth concepts, permission-aware retrieval and context preparation.

Data, APIs, Identity & Authorization

Design and implement the data and access boundaries behind an application: relational models, Prisma and PostgreSQL, APIs, integration contracts, authentication, RBAC, tenant-aware authorization and application-level data flows.

Platform Modernization & Replatforming

Existing systems do not always need to be replaced. Architecture can often be improved incrementally through clearer boundaries, framework modernization, database and API restructuring, deployment cleanup, performance work and migration planning.

Structured Publishing, Multilingual Systems & Technical SEO

Content architecture belongs to the platform. Multilingual routing, canonical URLs, hreflang, sitemaps, structured data, internal linking, EditorJS content models and semantic taxonomies can be designed together with the underlying application rather than added after development.

Infrastructure & Production Operations

A system is not finished when the code compiles. Deployment, Linux runtime behavior, reverse proxies, NGINX, process management, networking, logging, backups, migrations and troubleshooting are part of the production architecture.

Technical Delivery & Project Structure

Architecture and implementation can be connected with requirements, milestones, risks, acceptance criteria, backlog structure, validation and delivery documentation. IPMA® Level D and Professional Scrum Master I provide formal methods that complement the engineering work.

From Focused Task to Complete Product Path

An engagement does not have to cover every layer. It can start with an architecture review, a difficult backend component, a RAG pipeline, a frontend feature, an AI integration, a migration or an infrastructure problem. The difference is that decisions are made with the surrounding product and platform architecture in view.

Implementation Evidence

The technical approach is demonstrated through systems I design, build and operate: stajic.de, loving.rocks, figure.rocks, bazify.com, Aaasaasa AI Client, SenseFlow and related architecture and implementation work. These projects expose different problems across publishing, multilingual systems, AI integration, permissions, data, infrastructure and product delivery.

Technical Foundation

Current work commonly combines TypeScript, JavaScript, Nuxt, Vue, Node.js, Nitro, Prisma, PostgreSQL, Electron, Linux, NGINX, Docker, REST APIs, OpenAI integrations and local LLM runtimes including Ollama. Additional technologies are selected according to the system rather than used as fixed requirements.