← Portfolio Download PDF

Warre Snaet

Junior Backend & Applied AI Engineer

Vilvoorde, Belgium +32 474 82 94 93 warresnaet@icloud.com snaetwarre.github.io/My-Portofolio github.com/SnaetWarre linkedin.com/in/warre-snaet-272354370

Profile

AI Engineering graduate with distinction who builds dependable Python backends, agent infrastructure, applied AI systems, and data pipelines. Experienced in taking unfamiliar problems from raw data to validated software, cloud ML workflows, APIs, and deployable applications. Looking for a junior backend, applied AI, or MLOps role where I can contribute and grow with experienced engineers.

Technical experience

AI & Software Engineering Intern

2Ai IPCA · Barcelos, Portugal

Feb–Jun 2026

  • Built a Python pipeline that reads echocardiographic DICOM studies, detects burned-in measurement regions, extracts values with OCR and image processing, and exports structured JSON.
  • Designed resumable batch processing, validation queues, review tooling, ECG trace extraction, and cross-platform handoff documentation.
  • Developed a source-video matcher with OpenCV ECC alignment and DICOM metadata, reducing one representative run from roughly 21 hours to 23 minutes.

Software Engineer · Paid Client Project

Apolloon 24-Hour Run System · Belgium

2026

  • Built a local-first race operations platform with React, TypeScript, Express, Electron, SQLite, tRPC, and Socket.IO.
  • Implemented operator workflows, live displays, imports and exports, automatic LAN discovery, realtime synchronisation, and replicated hot-standby failover for unreliable event conditions.

Selected projects

Offline Semi-Supervised AI in Rust

Built an end-to-end Burn/Rust plant-disease classifier. The saved SSL checkpoint reached 94.90% held-out top-1 accuracy; packaged weights were about 916 KB and the offline application about 26 MB, with inference tested on an iPhone 12.

Financial AI Agent

In a two-person academic project, owned backend architecture, orchestration, MCP integration, persistence, containers, tests, and most frontend integration across a six-service application.