# Patricio Cerda-Mardini

> ML SWE @ Google

Location: San Francisco, California, United States
Profile: https://flows.cv/patricio

Machine Learning Research Engineer with 5+ years of experience in early stage startups & big tech.

Currently working on LLM agent reasoning and tooling for conversational analytics services at GCP. Previously, I worked on open source projects at the intersection between MLOps and AutoML. 

Research interests:
    - Large language models (agents, evals)
    - Uncertainty estimation (model calibration)
    - Machine translation
    - Cognitive robotics (navigation)
    - Recommender systems
    - AutoML
    - MLOps
    - Time series forecasting

## Work Experience
### Software Engineer @ Google
Jan 2024 – Present | San Francisco, California, United States

### Senior Machine Learning Research Engineer @ MindsDB
Jan 2020 – Jan 2024 | San Francisco, California, United States
Core maintainer in MindsDB, a popular MLOps open source project. 

Some of my contributions: 

1) co-creator and maintainer of Lightwood, a PyTorch-based declarative AutoML modelling framework for tabular use cases

2) co-authored a full paper at CoLM 2024 on LLM evals, and two posters at PyTorch Ecosystem Day 2021

3) designed and implemented base abstractions to integrate with 200+ databases and ML frameworks

4) created some of the most used ML integrations (OpenAI, LangChain, Lightwood) totalling over 25k deployed models

5) lead engineer on several features: forecasting, finetuning, LLM agents, model evaluation

### Machine Learning Engineer @ Dictuc
Jan 2019 – Jan 2020 | Santiago Metropolitan Region, Chile
Worked part-time on a CNN based system for real estate appraisal, using a variety of multi-modal input features.

### Research Assistant @ IALab PUC
Jan 2018 – Jan 2020 | Santiago, Región Metropolitana de Santiago, Chile
Cognitive Robotics group member, co-supervised by professors Álvaro Soto & Hans Lobel

### Teaching Assistant @ Pontificia Universidad Católica de Chile
Jan 2016 – Jan 2019 | Provincia de Santiago, Chile
Recommender Systems, Mobile Robotics, Fundamentals of Robotics, Programming as a Tool for Engineering.

### Intern, Computer Vision Research @ Metric Arts
Jan 2018 – Jan 2019 | Provincia de Santiago, Chile
Worked on a variety of topics: person re-identification in video streams using YOLO, license plate detection with small ConvNets, face classification using triplet loss, and content-based fashion recommendation systems using ResNets.


## Education
### Master in Engineering in Computer Science School, Artificial Intelligence
Pontificia Universidad Católica de Chile

### Licentiate in Engineering Sciences in Robotics Major, Computer Science Minor
Pontificia Universidad Católica de Chile


## Contact & Social
- LinkedIn: https://linkedin.com/in/paxcema
- Portfolio: https://paxcema.github.io

---
Source: https://flows.cv/patricio
JSON Resume: https://flows.cv/patricio/resume.json
Last updated: 2026-03-29