Building production-grade AI systems that deliver measurable impact. Specialized in semantic search, agent orchestration, and large-scale embedding pipelines.
AI Engineer with strong Software Engineering foundations and hands-on experience designing and deploying scalable, production-grade AI systems in the e-commerce domain.
Specialized in semantic search, agent orchestration, and large-scale embedding pipelines using modern LLM frameworks and vector databases.
I focus on building reliable, high-performance AI systems that enhance user experience, ensure operational robustness, and drive real business outcomes.
Stateful email agent from LangChain Academy with persistent memory and human-in-the-loop mechanisms. Full deployment on LangGraph Platform with LLM-as-a-judge evaluation and traceability via LangSmith.
Conversational AI agent for real-time news fact-checking. Implements a RAG workflow with GPT-4 + SerpAPI for multi-source verification, delivering unbiased summaries with a truthfulness indicator. Deployed on Azure Container Apps.
View on GitHub →Multi-agent simulation environment for RoboCup-inspired soccer scenarios. Implemented and fine-tuned Reinforcement Learning algorithms (PPO, MA-POCA) achieving a 35% improvement in team precision and coordination. Evaluated with TensorBoard metrics.
5-year degree completed in 4.5 years · Average grade: 8.05 / 10
OpenAI API · LangChain · Vector Databases · Embeddings · LLMOps
Deep Neural Networks · CNNs · Reinforcement Learning · Optimization
Why single queries fail in complex RAG systems — and how generating multiple reformulated queries from different perspectives dramatically improves retrieval quality and LLM response accuracy.
Read on LinkedIn →Intensive program at Universidad de Buenos Aires covering Diffusion Models, LLM training & scalability, advanced evaluation metrics, agent-based security automation, and optimization methods.
View post →Built VeriBot with teammates — a RAG-powered conversational agent for real-time news fact-checking using Azure Bot Service, GPT-4, and cross-source truthfulness analysis. Open source on GitHub.
View project →Open to new opportunities, collaborations, and interesting problems in AI engineering. Let's build something impactful together.