Selected work
These case studies focus on my contribution, the constraints that shaped the architecture, and what changed after the work shipped.
Case studyI joined Krafted AI as the founding AI engineer and built the product from scratch - from product strategy and system architecture to the Shopify app, AI workflows, infrastructure, and production deployment. It grew from zero to $18,000 MRR with 530 active users.
Read the decision trail
Case studyI built MintFit from the coach workflow outward, connecting assessments, client history, program design, and retrieval-augmented AI to the studio's own Thrive50+ methodology.
Read the decision trail
Case studyI built a direct Bézier control-point detection pipeline for noisy maritime imagery, combining curvature-aware RANSAC, non-linear refinement, occlusion recovery, and a new 5,000-image benchmark.
Read the decision trail
Case studyI engineered core parts of a global pharmaceutical publishing platform where localisation, medical review, market eligibility, accessibility, and editorial speed had to coexist in one content model.
Read the decision trail
Case studyI engineered core parts of Germany's regulated incident-reporting platform, where identity, deadlines, encryption, and every state transition carried a security or legal consequence.
Read the decision trail
Case studyI engineered the mobile and security layers of a non-custodial wallet, where key management, device protection, and a fluid interface had to coexist across more than 50 production screens.
Read the decision trail
Case studyI built the context, routing, and escalation system behind a multilingual WhatsApp assistant for more than 60 properties. Correctly identifying the guest and property mattered more than generating a fluent reply.
Read the decision trail
Case studyI engineered the structured evaluation pipeline behind a candidate-screening platform that combined video analysis, adaptive assessment, and interview intelligence without turning model output into an unexplained hiring decision.
Read the decision trail
Case studyI redesigned retrieval across two million ADAC documents, separating tables, legal guidance, vehicle tests, and editorial content into strategies that raised benchmark precision from 61% to 84%.
Read the decision trail
Case studyI built a reproducible research pipeline for eight million Scopus records, resolving researcher identities, detecting migration, and correcting gender-inference errors that commercial APIs reported with high confidence.
Read the decision trail
Case studyI proved and delivered a mid-project CMS migration that replaced a slow in-house workflow with a shared Storyblok component model across 37 storefronts and six languages.
Read the decision trail