Leonardo Rocha
AI/ML Engineer with 12+ years in Data Science and 5+ years building and deploying intelligent systems.
Specialized in Graph ML, NLP, and Generative AI — including Agents and RAG architectures.
A hands-on engineer with entrepreneurial vision, leadership experience, and a proven ability to transform state-of-the-art AI into scalable real-world systems.
Key Highlights
Led the full-cycle design and deployment of an AI-powered robotic agent automating legal workflows across 10,000+ monthly cases in a region of 9M inhabitants. Combined LLMs, rule engines, and RPA for decision automation, achieving >95% success rate according to human validation.
Designed and implemented adaptive learning and recommendation engines using semantic embeddings and graph-based modeling, powering a national ICT training platform and improving learner engagement and completion rates.
Built a multi-agent conversational platform that simulates expert panels (CTO, PM, VC) to mentor startup founders. Agents collaborate through LangGraph orchestration and RAG-based contextual grounding, ensuring coherent, knowledge-driven dialogue. Deployed with FastAPI and Google Cloud Run for scalable interaction.
Advised a US fintech startup on RL-based customer personalization and infrastructure design. Modeled uplift-optimization strategies and evaluated cost–benefit scenarios for integration with Stone and C6 Bank in Brazil.
Developed an intelligent cybersecurity module using graph-based machine learning to detect malicious websites through relational analysis of URLs, domains, and host networks. Results were peer-reviewed and published at the 37th International Conference on Advanced Information Networking and Applications (AINA 2023).
Founded SofIA Tech, bridging state-of-the-art AI research with market applications through consulting, automation projects, and enterprise AI training. Delivered multiple solutions using LLMs, RAG, and intelligent automation frameworks.
Founded the Graph and Computational Intelligence Lab (LAGIC) and coordinated the IncubaUECE incubator, fostering applied AI research, technology transfer, and entrepreneurship in deep-tech ventures across Ceará.
About
Machine Learning Engineer specialized in Graph ML, LLM agents, embeddings, Retrieval-Augmented Generation (RAG), and automation of real-world workflows. Proven ability to lead projects that deliver measurable efficiency gains with startup-level execution.
Experience
Led design and deployment of an AI-powered robotic agent automating legal workflows. Achieved >90% accuracy in key tasks, 98.7% dispatch approval without human revision, and >40% productivity gains.
Advised on reinforcement learning system for customer personalization. Defined technical architecture, mapped requirements, and evaluated infrastructure, costs, and risks to inform product roadmap decisions.
Led design and implementation of adaptive learning features for a national ICT training platform (5,000+ students). Built recommendation engine with 95% success rate in human-supervised evaluations.
Lead R&D in Data Science and AI. Supervised 20+ projects, published 20+ peer-reviewed papers, coordinated IncubaUECE, and founded the Graph and Computational Intelligence Lab (LAGIC).
Skills
Education
Get in Touch
Prefer email? Reach me directly.