Designed and built a serverless, asynchronous job-processing API on Azure Functions. Structured the solution using Clean Layered Architecture (Domain, Application, Infrastructure, Functions) with Dependency Inversion so Azure-specific code is isolated and swappable.

Implemented an HTTP-triggered intake that persists a job and enqueues work, plus a queue-triggered background function driving a status state machine (Queued to Processing to Completed/Failed).

Built repository and messaging abstractions over Azure Table Storage and Azure Queue Storage, provisioned the full Azure footprint declaratively in Bicep, and covered the application layer with xUnit unit tests using in-memory fakes.

Working as a IT consultant, currently developing an internal application.

Long-term technical contributor in a restricted, offline environment with end-to-end responsibility for archiving sensitive data, primarily medical records, from decommissioned systems. Collaborated with archivists and healthcare experts to scope preservation needs and design the technical specifications.

Led the region's two largest archival projects (Melior for open care, Orbit for surgical data), each spanning 1 to 1.5 years and processing multi-terabyte databases. Reverse-engineered data models from undocumented databases through stored procedure tracing and semantic analysis.

Sole developer of the eArkiv AdminInterface, a three-tier ASP.NET MVC 5 application (DDD) used by ~30 archivists. Built its access control from scratch on IIS Windows Integrated Authentication, validating each user's RSID against a DB allow-list and injecting a role-based security principal. Upgraded the app from .NET Framework 4.5.2 to 4.8.1 and jQuery 1.x to 3.x, and cut SQL that previously ran 1-2h down to seconds.

Developed multi-threaded BLOB extraction, streaming XML parsers, and high-performance SQL with batching, strategic indexing, and staging tables, including the stored procedures that measured storage consumption across the ~5 TB production database to distribute operating costs per care unit. Maintained a GDPR-relevant Web API for personnummer correction, and contributed to HL7 FHIR and SIARD preservation standards.

Worked on an AI support solution designed to analyze historical support tickets and provide categorization and solution suggestions to support staff.

Designed and deployed the complete Azure backend, configuring Linux VMs with Nginx reverse proxies and automated CI/CD pipelines via Azure DevOps, with a dedicated managed identity scoped to least-privilege RBAC roles. Implemented automated scaling using Azure Automation to minimize costs.

Built the data processing pipeline using CosmosDB and MongoDB, including embedding generation via local AI models (OLLAMA) and a custom vector-search pipeline against MongoDB's native client with proper document structuring, chunking, and cleaning.

Removed a major bottleneck by replacing an O(n) per-request list scan with a keyed in-memory lookup (O(1)), and fixed a hallucination issue by strengthening prompt engineering and adding a null/empty-data guard so the system would not fabricate answers when data was missing.

Contributed to application design and data modeling for reporting and search functionality as part of a small development team. Participated in system design discussions including database normalization and data modeling.

Developed and maintained the public website using Episerver CMS (today Optimizely). Worked with frontend and backend web development, feature implementation, and content handling in an agile team.

Developed interactive data visualizations in JavaScript to model and graphically represent patient rehabilitation metrics, enabling clinical staff to track progress trends.

Worked as IT consultant with various projects and customers.

Completed an accelerated 12-week C#/.NET training program through Academic Work, then worked as an integrated consultant at AddPro AB (later ITM8) on client projects. Got bought over to AddPro after completed contract.

Part of various projects ranging from GDPR system evaluation to working with GE Machine Execution System products such as Plant Applications with an MS SQL Server backend. Contributed to internal web development and supported UI implementation by reviewing and documenting HTML structures.

Led Slow and Obsolete Goods (SLOBS) operations from the Quality Department, solving IT and goods-related issues, implementing improved routines, and ensuring SOX compliance. Recovered goods worth several million SEK back into sale and production. Performed risk assessments supporting go/no-go decisions for raw, semi-, and finished goods. Served as Lab Operation Manager during summer, overseeing tens of thousands of tons of raw material. Early coordination in strategic warehouse consolidation enabled ~15% of the order list to be handled ahead of schedule.