Résumé

Datachecks Enterprise: AI agents for enterprise data migration, at a scale where the interface is the hard part.

• Architected multi-agent systems with persistent memory (Mem0), RAG pipelines and agent orchestration for automated dataset discovery and synthetic data generation.
• Led a four-person frontend team on Next.js, shipping RBAC, SAML SSO, multi-workspace support, validations, cross-database comparison, lineage and profiling.
• Powered the on-premise deployment for the UAE Central Bank migration: 2× faster than the manual QA benchmark, saving around 60% of the time.
• Merged 350+ PRs across four repositories, spanning the frontend and the agent systems behind it.

Datachecks SDK: the open-source validation engine underneath the platform.

• Shipped production features on a database validation platform used by enterprise clients including John Deere, on datasets of 10M+ rows.
• Delivered 8+ features across frontend and backend, cutting validation set-up time by roughly 40%.
• Built the React frontend for schema exploration and validation results, cutting new-user onboarding time by around 30%.

• Built the end-to-end MVP: real-time speech-to-text with LLM summarisation into structured SOAP notes, with clinician review before anything is saved.
• Reached ~90% transcription and summarisation accuracy in live pilots, with an estimated 50–60% cut in documentation time.