AI Platform Engineer

I own the architecture of modular, configurable AI agents in partnership with the CTO, taking systems from design through production deployment. I use Pydantic-validated structured outputs, Docker deployments, and Ray-based horizontal scaling to build reliable AI services.

I build RAG systems and evaluation harnesses that measure agent accuracy across LLM providers and support data-driven model selection. I integrated OpenAI, Mistral, Groq, and Qwen/Kimi models to optimize quality and cost, achieved 99% accuracy for document-processing workflows, and automated more than 80% of brokers' repetitive processes. I also mentor junior team members in agent design, pipeline development, clean code, and modular architecture.

I deployed and optimized AI models for audio-to-text and text-to-text transformations using Hugging Face tools, connecting speech recognition and natural language processing workflows.

I designed and fine-tuned YOLO-based object-detection models for real-time image recognition and tracking. I also engineered web-scraping systems to extract, clean, and analyze data from diverse online sources, producing enriched datasets for data-driven decision-making.

I developed a supervised machine-learning predictive ranking model to optimize Google search-page rankings, strengthening SEO strategy and increasing organic traffic and revenue.

I analyzed millions of records to identify actionable patterns and trends for content strategy and marketing campaigns. I also collected and integrated multi-source data through REST APIs and advanced web scraping, improving dataset accuracy and completeness by 40%.