I build AI systems in production today - and move steadily toward performance, local inference, and systems-level engineering.
Backend & AI engineer with 7+ years of experience building production systems. I design reliable AI backends, RAG pipelines, APIs, retrieval workflows, and evaluation layers that hold up in real environments. My trajectory is deliberate: from applied AI delivery into fine-tuning, performance engineering, local inference, and C++-oriented systems work. I care about architecture, latency, correctness, and the deeper technical foundations behind modern AI systems. I don't separate research from engineering. Production is where research meets reality - and that's exactly where I want to be.
Production AI systems
Built a RAG pipeline that reduced document retrieval latency by 40% while maintaining recall quality across a production knowledge base.
Performance as a discipline
Designed an eval layer that surfaced hallucination cases before production. Latency, observability, and correctness treated as engineering constraints - not afterthoughts.
From builder to researcher
Production engineering is my laboratory - it gives me the constraints, failures, and load patterns that research alone cannot. My trajectory moves from applied AI into the systems-level questions: how models behave under pressure, where inference breaks down, and what it actually takes to optimize at depth.