I build production AI systems — LLMs, RAG, and multimodal pipelines that actually ship.
AI/ML Engineer with experience designing, developing, and deploying production-ready AI applications powered by Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), multimodal AI, and machine learning. Experienced in building scalable AI services, fine-tuning transformer models using PEFT (LoRA and DoRA), and integrating NLP, computer vision, OCR, and speech technologies into production systems. Passionate about transforming cutting-edge AI research into reliable, scalable AI products through efficient inference, modern AI infrastructure, and backend engineering. Recognized with a Kaggle Silver Medal (Top 1.6%) in the AI Mathematical Olympiad and a national research grant for applied LLM research.
Comparative evaluation of PEFT methods on LLaMA models for subjective NLP tasks, across adapter ranks on the Chatbot Arena dataset. DoRA won on accuracy and human-preference alignment; LoRA on speed and memory — a practical map for resource-constrained fine-tuning.
LLM reasoning pipelines combining prompt engineering, test-time reasoning, answer verification, and efficient inference — solving Olympiad-level math with open-source models under strict compute constraints. Top 1.6% worldwide (64th of 4,138 teams).
Grant-funded research on improving the interpretability and efficiency of LLMs for automated medical diagnosis — toward transparent, reliable medical AI systems.
NLP-powered chatbot enabling context-aware analysis of network KPIs through natural-language interaction, paired with Prophet forecasting dashboards.
WhatsApp OTP multi-factor authentication integrated with an AI-based fraud detection system — data security, model integration, and digital risk prevention.
AI-driven climate engineering for hydroponic–aeroponic plant growth using IoT automation — awarded for originality, technical execution, and societal impact.
The amber meter under each badge reflects how often the skill appears across roles, projects, and awards in the CV — not a self-assessed score.
Open to AI/ML engineering roles and applied-LLM collaborations — from RAG systems to multimodal pipelines.
rivaldofr03@gmail.com