Dmytrii Shchadei

I help early-stage and scaling B2B AI companies turn vague technical/product ideas into credible prototypes, architectures, and production-ready systems. My strongest fit is where the problem is messy: AI agents, ML platforms, distributed systems, workflow-heavy products, compliance constraints, reliability, data pipelines, and high-performance user interfaces. I’ve spent 15+ years working across the AI stack — ML modeling and training, MLOps, distributed systems, reliability engineering, frontend architecture, and product-facing technical leadership. That breadth lets me connect product intent with engineering reality faster than narrow specialists. Previously, I led technical delivery at Meta on multimodal support diagnosis systems, LLM/state-machine/logic-programming agent frameworks, policy enforcement, guardrails, fraud, privacy, and abuse controls. Before that, I worked on ML systems for robotaxi fleets at Mercedes-Benz and Vay, large-scale production ML at AWS, and high-performance mapping platforms at HERE/Yandex-scale environments. I’m currently building Swobu — a local-first boundary layer that unbundles AI clients from LLM vendors. Open to selective Outside IR35 / 1099 contract, advisory, prototype sprint, and fractional CTO work in AI/ML, agentic systems, developer tools, infrastructure, and complex production platforms. Useful engagements: — 2–4 week AI prototype sprint — technical/product architecture audit — fractional CTO support for nontechnical or product-led founders — agentic system design and reliability review — ML platform / infrastructure / workflow system design — investor or customer demo hardening