I am an applied AI engineer and former CTO with more than 10 years of R&D and production engineering experience. I specialize in transformer, LLM, retrieval, ranking, and multimodal machine learning systems across text, audio, tabular, and clinical data.
I currently own a production agentic RAG pipeline and its fine-tuning roadmap. My work includes multi-step reasoning, cross-encoder reranking, section-aware document parsing, evaluation harnesses, and reliability improvements for LLM outputs.
I have hands-on experience with PEFT, LoRA, QLoRA, Llama-3, Mistral, Hugging Face, PyTorch, and TensorFlow. I balance research quality, scalability, and ROI when moving AI solutions from experimentation to production.
I have collaborated closely with academic researchers, clinicians, external clients, and national-security research teams. My projects include clinical risk prediction, audio stroke detection, legal semantic search, cybersecurity data analysis, and NLP classification systems.
As a former CTO and technical lead, I have led multidisciplinary teams and delivered end-to-end platforms. I am comfortable owning ambiguous problems from ideation through architecture, implementation, evaluation, and production deployment.
I also mentor machine learning and statistics practitioners, including PhD students. My background combines applied research leadership, engineering execution, and a strong focus on measurable model quality and system reliability.
