Backend Engineer.

Systems-oriented backend engineer with a track record of rapid domain acquisition and production-grade delivery across genomics, finance, legal AI, and enterprise knowledge management. Consistent architectural approach across all domains: clean separation of concerns, protocol-based extensibility, event-driven patterns, and comprehensive observability. Currently building enterprise knowledge infrastructure on Azure.

Featured projects

Built to ship.

Every project exhibits the same design principles: clean separation of concerns, protocol-based extensibility, event-driven patterns, and comprehensive observability.

Autopod Multi-Tenant Podcast Orchestrator

First Go code ever written — deployed as a multi-tenant production system with full test coverage and CI/CD.

  • Control plane / data plane split: Go orchestrator drives a 7-step pipeline by calling stateless Python/FastAPI endpoints with typed HTTP contracts mirrored across languages
  • Crash-resilient job state machine with SQLite persistence — every state transition writes to disk before proceeding, GetActiveJobs() resumes all in-flight jobs on startup
  • Multi-tenancy by construction: per-customer watch directories, age-encrypted secrets decrypted at runtime, namespaced cloud storage on Cloudflare R2
  • 21 Go tests + 36 Python tests, GitHub Actions CI with -race flag, Docker multi-stage builds

Memory Palace Graph-Based Semantic Memory System

Personal/learning project. Query language design, graph databases, and the specification pattern — all learned from scratch for this build.

  • Clean architecture with enforced layer boundaries: domain → infrastructure → services → api. Protocol-based interfaces for all external integrations
  • Extensible query API using discriminated union specification pattern: 12+ composable spec types generating both Python predicates and Cypher WHERE clauses
  • Type-safe Cypher query builder with state machine validating clause ordering at construction time
  • Integrated as MCP server enabling Claude.ai to use the system as persistent external memory

Grimoire (SETS) Enterprise Knowledge Management Platform

Enterprise evolution of Memory Palace, deployed on Azure at Wise. Same architectural DNA evolved for cloud infrastructure ownership and team-scale deployment.

  • Deployed on Azure Container Apps with Bicep IaC: Neo4j AuraDB, Azure Container Registry, auto-scaling (scale-to-zero dev, 1–3 replicas prod)
  • Full OAuth 2.0 authentication spanning three RFCs (7591, 8414, 9728) — Claude.ai MCP connector auto-negotiates auth, zero-config enterprise deployment
  • Pluggable connector architecture: Confluence space indexing, GitHub repo analysis, code complexity metrics — all queryable through unified specification-based JSON DSL
  • Enterprise features: JWT identity tracking, background job orchestration with APScheduler, liveness/readiness health probes

Sokrates IDR Legal Document Intelligence Platform

Solo-built full-stack enterprise platform (estimated 4,000–8,000+ development hours by independent review).

  • DDD with architectural boundaries enforced by import-linter contracts: domain → core → infrastructure → api. Repository pattern with protocol-based interfaces
  • Polyglot persistence: PostgreSQL (asyncpg + SQLAlchemy + Alembic), Neo4j knowledge graph, MinIO document storage, PGVector embeddings
  • OpenTelemetry with 13+ instrumentations, Logfire + Sentry + structlog for structured observability
  • Comprehensive testing: pytest-asyncio, factory-boy, Playwright E2E. Automatic TypeScript client generation from OpenAPI specs
By the numbers

Impact across domains.

0.42%

Forecast accuracy on 2.4B ISK cash flow volume

>10×

Compression ratio vs gzip for haplotype data

97%

Improvement in transaction reconciliation

4,000+

Solo development hours on Sokrates IDR

Career

Professional experience.

Oct 2024 — Present

Backend Engineer

Wise (Iceland)

Enterprise knowledge management and internal tooling. Designed and deployed the Grimoire/SETS platform on Azure infrastructure for team use. Working with Microsoft Business Central integrations and connector development.

Aug 2022 — Sep 2024

Data Scientist & Business Intelligence Lead

Travelshift

Complete overhaul of data infrastructure, financial process automation, and predictive modeling for Iceland's largest travel marketplace.

  • Kimball-dimensional star schema in Snowflake, ETL pipelines, Power BI reporting
  • Cash flow forecasting with 0.42% margin of error on 2.4B ISK volume (6-month horizon)
  • Automated payment reconciliation: 1-in-35 → <1-in-1,000 unreconcilable transactions
  • Reduced booking department from 7 contractors to 2 while improving performance

May 2021 — Aug 2022

Data Analyst & Product Owner

Alfreð Atvinnuleit

Established BI environment with real-time dashboards. Product owner for gig economy freelancing platform. Authored grant proposal securing 30M ISK funding from Rannís.

2015 — 2017

Statistician & Bioinformatician

deCODE Genetics (Amgen)

Designed novel haplotype compression algorithm achieving >10× compression ratio versus gzip — enabled loading entire chromosomes into memory. Processed and analyzed large genetic datasets.

2018 — 2021

Instructor, Department of Computer Science

Reykjavík University

Taught while completing MSc: Programming, Data Structures, Calculus & Statistics, Discrete Mathematics II.

2013 — 2017

Founder

Taurus Supplements

Founded and operated a specialized supplement product line. End-to-end ownership: product development, manufacturing coordination, marketing, sales, and daily operations.

Technical skills

Tools and technologies.

Deep experience across the full stack, from architecture to observability.

Languages

Go (building proficiency — see Autopod), Python (expert), TypeScript/React, SQL, R, DAX/M

Backend & APIs

FastAPI, RESTful API design, OpenAPI specification, automatic client generation (Orval), WebSockets, Pydantic, dependency injection

Databases & Storage

PostgreSQL (asyncpg, SQLAlchemy, Alembic), Neo4j (Cypher, graph modeling), Snowflake, MinIO (S3-compatible), PGVector, SQLite

Cloud & Infrastructure

Azure (Container Apps, ACR, Bicep IaC, Databricks, Log Analytics), Docker (multi-stage builds, non-root), Cloudflare (R2, Tunnel), GitHub Actions CI/CD

Architecture & Patterns

Domain-Driven Design, clean/hexagonal architecture, specification pattern, repository pattern, event-driven design, control plane/data plane, Kimball-dimensional modeling

Observability & Quality

OpenTelemetry (13+ instrumentations), Logfire, Sentry, structlog, Ruff, basedpyright, Biome, import-linter, pre-commit hooks

Security

OAuth 2.0 (RFC 7591/8414/9728), JWT, bcrypt, age encryption, Google Cloud OAuth, CORS, non-root containers

AI & ML

LLM integration (Anthropic, OpenAI, Google Gemini), RAG systems, vector embeddings (Voyage AI), pydantic-ai, Claude Code, DBSCAN clustering

Data Engineering

ETL pipeline design, Microsoft Fabric, Power BI, PySpark, Celery distributed task queues, star schema design, data warehouse architecture

Education

Academic background.

MSc Computer Science

Reykjavík University · 2017 — 2021

Artificial Intelligence and Data Science. Thesis: Recommendation Systems.

BSc Mathematics

University of Iceland · 2011 — 2016

Computational Mathematics and Computer Science.

Let's connect.

Looking for a backend engineer who ships production systems across domains? I'm always open to interesting conversations.