Skip to main content
~/mario-de-lucas
Mario De Lucas Garcia

// mario de lucas garcia - ai engineer, new york

Practical AI systems, built to make life better.

By day I lead AI automation at FlatironDragados. Before that: fine-tuning LLMs at a startup and shipping cross-platform apps. Curious by nature, pragmatic by default - I like AI that quietly improves someone's day.

run fit_check(your_role) -> ./experience

experience.log

Experience

2024 - now

Software Engineer - ML & AI Specialist

FlatironDragados - New York

Spearheading AI initiatives across business processes: finding automation opportunities, turning raw data into decisions, and building analytics systems people actually use.

  • python
  • ml
  • automation
  • power-bi

2023

Machine Learning Engineer

Zatori - Madrid

Engineered the dataflow for training and fine-tuning company-specific language models. Shipped a support chatbot that cut customer-service costs without cutting quality.

  • llm fine-tuning
  • langchain
  • pytorch

2021 - 2023

Full-Stack Engineer

WE Community - Madrid

Built cross-platform mobile apps in React Native and a recommendation system that made the product feel personal - and customers stick around.

  • react-native
  • node
  • recsys

2020 - 2022

Research Assistant

Ontology Engineering Group - UPM, Madrid

Knowledge representation for genuinely messy domains, including VAT law and logistics. Led the semantic model for the SIMPLE project.

  • ontologies
  • semantic-web

education: M.S. artificial intelligence, 4.0 GPA (Illinois Tech) - M.S. computer science - B.S. computer science + B.S. business admin (Madrid)

honest_fit_assessment

Pick the role you are hiring for.

Your time matters. Choose a role and get the honest answer: where I am strong, where I am not, and whether I am worth the call.

ml-engineer fit assessment

88% fit

Strong fit. This is what I do daily.

strengths

  • Production ML at 3 companies - recommendation systems, chatbots, automation
  • PyTorch, TensorFlow, and scikit-learn as daily drivers
  • M.S. in AI, 4.0 GPA - theory that backs the practice

honest gaps

  • Petabyte-scale MLOps (Kubeflow, feature stores) - learning it, not yet expert

assessment_time: ~20s - phone_screens_saved: 1 - honesty: 100%

// beyond_the_code

Beyond the code

Away from the keyboard: running along the Hudson, early gym sessions, and soccer or tennis on the weekend. The thread through everything I build: AI should make someone's life measurably better. That is the filter.

email linkedin github