Barcelona, Catalonia, Spain
I build AI systems that ship. ๐๐บ๐ฎ๐ถ๐น: [email protected] ๐ช๐ฒ๐ฏ๐๐ถ๐๐ฒ: https://danielemingolla.com/ At OpenNebula I contribute to OneAI, a CLI that deploys LLM-serving VMs in one command, and built the open-source MCP server (one-mcp) that cuts cloud management time via natural language. Speaker at FOSDEM 2026 and Open Source Experience 2025. ๐ฃ๐ฟ๐ฒ๐๐ถ๐ผ๐๐๐น๐: ML recommendation systems at Glovo (โฌ4M projected annual impact), game analytics and ETL at SocialPoint, and data engineering at Amazon. I also founded TechPuente (SaaS + growth via TikTok and SEO) and host the yLearner Podcast on career growth and business. ๐ง๐ฒ๐ฐ๐ต๐ป๐ถ๐ฐ๐ฎ๐น: LLM fine-tuning, vLLM, RAG, MCP, LangChain ยท PyTorch, Hugging Face ยท AWS, Docker, Kubernetes, Spark, FastAPI, MLOps ยท Causal inference, A/B testing. Open to technical collaborations and speaking. Reach out: [email protected]
โข Implemented fine-tuning workflows and automated GPU scheduling (PCI passthrough for H100/A100/V100) for the OneAI CLI, enabling one-command LLM deployment across OpenNebulaโs 5,000+ cloud installations. โข Built a dependency-free Time Series Forecasting SDK (NumPy/Polars) using Gradient Boosting and Fourier Transforms for predictive VM scaling, reducing cloud operational costs by 15%. โข Authored and released one-mcp, an open-source MCP server connecting LLMs to cloud infrastructure, enabling natural language DevOps and cutting cloud management time by 40%. โข Speaker at FOSDEM 2026 (Brussels) and Open Source Experience 2025 (Paris) on AI-assisted cloud management and the OneAI architecture.
โข Designed and analyzed A/B tests across multiple domains (UX, email, promos), collaborating with PMs and SWEs on experiment design, implementation, and statistical analysis. โข Worked on the recommendation system for in-app ranking and personalized communications (email/push), orchestrating implementation across teams with projected โฌ4M annual GMV impact. โข Validated EU discount policy impact through causal inference, proving fraud reduction (99.6%) without business degradation, enabling policy rollout across operating markets.
โข Analyzed arena mode matchmaking data to identify engagement bottlenecks, driving game design improvements that increased ARPU by 5%. โข Developed genetic algorithm for game balancing automation, reducing game designersโ manual workload by 70% while preserving gameplay quality. โข Architected ETL pipeline for real-time game metrics across 4 teams, reducing weekly reporting time from hours to minutes.
โข Optimized Spark job performance through code refactoring and resource optimization, reducing AWS costs by 20%. โข Developed ETL pipelines in SQL to power QuickSight dashboards, enabling data-driven operations across warehousing teams. โข Collaborated with Research Scientists on temporal graph neural networks to model warehouse dependencies for improved order forecasting across 1,000+ facilities.
โข Produced and recorded 10 AI-focused podcast episodes, demonstrating expertise in audience research, guest coordination, and impactful question preparation. โข Contributed to the sales of approximately 20 masterclasses, generating over 6,000โฌ in revenue for ProfessionAI.