Multi-Agent LLM Pipelines
Architected multi-agent pipelines that integrate LLMs with external APIs and databases — orchestration, tool-use, and decision layers composed into reliable, production-grade automation.
AI SOFTWARE ARCHITECT
I architect multi-agent AI systems that turn complex business problems into production-ready, autonomous products — built on a decade of full-stack and product engineering.
TECHNICAL EXPERTISE
A decade of building robust systems with battle-tested technologies
Multi-agent pipelines, LLM orchestration, RAG, and tool-use integrations with external APIs & databases
Claude Code & Codex as primary build tools — agent-driven development end-to-end
AI tooling, automation, APIs, and data pipelines
Enterprise-grade applications, microservices, and complex systems
RESTful APIs, backend architecture, and production services
Modern frontend development with TypeScript
Deep understanding of complex systems and algorithmic thinking
Coordinating LLM agents, tool-use, and decision layers into reliable pipelines
Grounding models in external APIs, databases, and retrieval
100k+ lines of code in autonomous systems
Self-optimizing orchestration and decision engines
Telemetry pipelines and performance tracking
Data-informed architecture and optimization
SYSTEMS & PROJECTS
Complex systems built with scalability, performance, and maintainability in mind
Architected multi-agent pipelines that integrate LLMs with external APIs and databases — orchestration, tool-use, and decision layers composed into reliable, production-grade automation.
Led all technical architecture, stack decisions, and system design for an AI SaaS product. Built and shipped core product features end-to-end using Python, modern APIs, and AI tooling; owned the engineering roadmap and scalability strategy.
A self-hosted system of autonomous agents that run my own operations — an LLM-maintained knowledge base, scheduled research and reviews, and ops automation. Dogfooding agentic architecture daily on a self-managed stack.
Built a fully autonomous system running large-scale automated workflows. Features self-optimizing decision engines, comprehensive telemetry, and real-time monitoring across 100k+ lines of code.
Designed and implemented dashboards with live database connections, real-time analytics, and dynamic data visualization. Features pattern recognition and automated alerting systems.
ENGINEER MINDSET
The mindset and approach that drives impactful engineering
Proactively identifying opportunities, taking initiative, and demonstrating ownership. Founder of podcast 'Op Papier', always focused on value creation.
Fast learner who thrives in different workflows and project environments. Hands-on mentality with ability to quickly master new technologies and domains.
Strong analytical capabilities combined with deep understanding of how complex systems interact. Ability to see both the forest and the trees.
Deep reflection and self-improvement mindset. Not afraid to experiment or fail, but focused on learning and iterating towards better solutions.
Natural instinct for automation and efficiency. Constantly seeking ways to streamline workflows and eliminate bottlenecks.
Creating trust between stakeholders and development teams by translating technical complexity into understandable, valuable solutions.
"A good leader is someone who doesn't do everything themselves, but ensures the team can bring out the best in themselves."
— Personal Philosophy
PROFESSIONAL JOURNEY
Building products and systems across multiple industries
Architecting multi-agent pipelines that integrate LLMs with external APIs and databases — designing production-ready, autonomous AI systems end-to-end.
Led all technical architecture, stack decisions, and system design for an AI SaaS product. Built and shipped core features end-to-end with Python, modern APIs, and AI tooling, and owned the engineering roadmap and scalability strategy.
A year travelling through Indonesia, Vietnam, Thailand and Australia.
Optimised the IAM testing pipeline for a faster, more reliable workflow, and translated complex client requirements into concrete system architectures and delivery roadmaps.
Owned backlog prioritisation, scope control, and stakeholder alignment from concept to delivery. Shipped technically complex features using Next.js and Laravel within Agile/Scrum cycles.
Designed and built backend services and APIs (Java Spring Boot) and owned the product roadmap end-to-end. Managed a large-scale automation platform including telemetry pipelines tracking 500k+ events/day.
Led projects with responsibility for backlog, scope management, and stakeholder communication. Implemented Agile methodologies effectively.
Bridged business requirements and technical implementation. Focused on translating complex customer needs into actionable development tasks.
Hogeschool Utrecht
Graduated with Merit · Feb 2025
Practice-oriented HBO-ICT route specialised in Business & IT. Worked in multidisciplinary teams on real-world challenges.
Coursera (Google)
Sep 2022 — Feb 2023
Digital marketing fundamentals and strategy.
GET IN TOUCH
Looking for challenging technical projects where I can make a real impact. Open to discuss how I can contribute to your team.
[email protected]
@Tjonna Koster
+31 6 18 52 26 55
Nijkerkerveen, Netherlands