On-prem AI & automated review
Internal model serving on NVIDIA GPUs with open-weight LLMs, linking AI-assisted quality checks straight into engineering workflows.
Software and systems engineer based in Stockholm. Computer Engineering at KTH, now studying Innovation, Defence and Security at the Swedish Defence University.
I work on backend services, infrastructure, delivery automation and applied AI. Most of that work has been in defence, enterprise software and telecom.

Right now I'm at Avioniq, working on software and engineering infrastructure in the defence aviation space.
Before that: concurrent Go APIs and payment flows, generative-AI backends with Kafka and vector databases, and keeping national telecom infrastructure running from Tele2's network operations centre. It started with Combat Boat 90 engines in the Swedish Armed Forces, which taught me what "mission-critical" actually means.
Software / Systems Engineer · Kista
Software and infrastructure engineering for defence aviation, working across application development, automation and applied AI with multidisciplinary teams.
Degree Project · AI-assisted code review
Investigated AI-assisted code review against human expertise in enterprise development, and the quality and security trade-offs of putting it in CI/CD.
Backend Engineer
Concurrent REST APIs in Go on PostgreSQL and MongoDB. Hardened payment flows with better error handling and production-readiness reviews.
Backend Developer / Scrum Master
Generative-AI services in FastAPI, event-driven RAG pipelines over Kafka and vector databases, Kubernetes deployments, and ran the team's ceremonies as Scrum Master.
Network Operations Center Engineer
Monitored and troubleshot national telecom infrastructure under strict availability and SLA requirements; incident response and coordination.
Amphibious Systems Mechanic / Engineering Officer
Kept mission-critical systems on Combat Boat 90 running, with structured troubleshooting under time pressure.
Internal model serving on NVIDIA GPUs with open-weight LLMs, linking AI-assisted quality checks straight into engineering workflows.
Retrieval-augmented document search with semantic retrieval and custom chunking.
An API for embedding-based similarity search.
Event-driven automation with retry handling and workflow triggers.
Real-time multiplayer game engine with a custom network sync layer.
Real-time projectile detection (YOLO11) feeding a C++ physics engine that simulates trajectories. In development.
Model selection, token streaming and Markdown rendering.
Browser-based image processing with computer vision.
Linux, Docker, Kubernetes, Ansible, GitLab CI, GitHub Actions, GPU infrastructure, observability
LLM deployment, AI developer tooling, RAG, vector search, Kafka, Qdrant, FAISS, PostgreSQL, MongoDB, Redis
Go, Python (FastAPI), Java, C, C++, Rust, TypeScript, REST, gRPC, OpenAPI, Spring Boot
Incident response, network operations, infrastructure security, secure delivery, reliability
Cisco CCNAv7: Enterprise Networking, Security & Automation · Google Cybersecurity Analyst
Swedish Defence University. Innovation, security strategy and technology integration.
KTH Royal Institute of Technology. Software development, distributed systems, computer engineering.