Francisco Ramos Pérez

MEng in Autonomous Systems, University of California Berkeley
MSc in Robotics and Automation, Technical University of Madrid
BSc in Robotics and Automation, University of Seville

I am currently a Master's student at UC Berkeley, specializing in Computer Sciences with a deep passion for Natural Language Processing (NLP), Computer Vision, and advanced control algorithms for autonomous systems. My journey in academia and research has been marked by significant involvement in various cutting-edge projects and contributions to the field.

My research in NLP has primarily focused on fine-tuning Large Language Models (LLMs) and leveraging them for diverse applications like text-to-code conversion, machine translation, and generative modeling. A noteworthy project I undertook was at the Sky Computing Lab at UC Berkeley, where I conducted correlational analysis of LLMs. This research aimed to enhance their efficiency and develop evaluation methods that do not rely on labeled data, pushing the boundaries of conventional NLP methodologies.

In the realm of Computer Vision, I have worked on the development of an autonomous agent that controls a car using perception input and behavior cloning in AI Racing Tech . This venture into automotive autonomy was further enriched by my time at the Advanced Center for Aerospace Technologies , where I implemented an object detection model in a drone-based detect-and-avoid system. My expertise also extends to agricultural technology, where I have utilized computer vision to analyze satellite images for feature extraction in agricultural crops.

My academic journey has been marked by co-authoring three publications in esteemed computer vision and robotics conferences, highlighting my contributions to these fields. Additionally, I was honored with an award from the University of Sevilla for the best final project among over 600 candidates. This project revolved around using neural networks to optimize the management of EV charging stations by predicting power demand and generation, showcasing my ability to apply advanced technologies in practical, real-world scenarios.

Further showcasing my leadership skills, I served as the project manager for the Autonomous Unmanned Aerial Vehicle System at UPM. Leading a dynamic team of 12 members, we focused on developing a drone for an Autonomous Drone Racing Competition, blending intricate technological elements with team coordination and project management expertise.

I am seaking out new opportunities to broaden my knowledge and develop my professional career. If you think I would be a good fit for your lab or company, feel free to reach out to me.

Research interest

As a passionate individual with a deep fascination for cutting-edge technologies, I am particularly drawn to the fields of NLP, Computer Vision, and Autonomous vehicles. These research areas have captivated my attention due to their potential to revolutionize various industries and improve the quality of life.


Industry Projects

Recommendation system to reduce customer frustration, incorporating a prediction system for waiting times and optimizing prices and distances
General Motors
January 2024 - May 2024
Berkeley, California
Development of Autonomous GoKart including Perception, Planning, and Control through AI
AI Racing Tech - ROAR: Robot Open Autonomous Racing
August 2023 - May 2024
Berkeley, California
End-to-End Self-Driving with Behavioral Cloning in CARLA Simulator
AI Racing Tech - ROAR: Robot Open Autonomous Racing
January 2024 - May 2024
Berkeley, California
Paper
Enhancing LLM training optimization algorithms through the application of generalization metrics and correlational analysis
Sky Computing Lab. University of California, Berkeley
August 2023 - December 2023
Berkeley, California

Research Projects

Autonomous Unmanned Aerial Vehicle System: Drone for Autonomous Drone Racing Competition.
Francisco Ramos Pérez et al.
Project Manager (12 team members)
September 2022 - June 2023
Madrid, Spain
Charge demand and renewable generation forecasting with Deep Learning: application to electric vehicle station optimization
Francisco Ramos Pérez, Francisco Rodríguez Rubio, and Carlos Vivas Venegas
September 2021 - July 2022
Seville, Spain
Code Text
Utilized High Altitude Platform Station (HAPS) and Computer Vision techniques to apply photogrammetry for crop monitoring, enhancing crop management efficiency and analysis accuracy
HIBA: Hub Iberia Agrotech
February 2022 - April 2022
Seville, Spain
Implementation and training of anomaly detection algorithms for road inspections using images acquired by drones.
Omicron - Intelligent Road Asset Management Platform
May 2022 - June 2022
Seville, Spain

Publications

Charge demand and renewable generation forecasting with Deep Learning: application to electric vehicle station optimization
Francisco Ramos-Pérez, Francisco Rodríguez Rubio, and Carlos Vivas Venegas
September 2021 - July 2022
Paper Poster
Benchmark on real-time long-range aircraft detection for safe RPAS operations
Víctor Alarcón, Pablo Santana, Francisco Ramos, Francisco Javier Pérez-Grau, Antidio Viguria, and Aníbal Ollero
February 2022 - November 2022
Paper
Optimized Operation of an Electric Vehicle Charging Station with Photovoltaic Support and Vehicle-to-Grid Implementation
Carlos Vivas, Francisco R. Rubio, Antonia López, and Francisco Ramos
September 2021 - July 2022
Paper

Course Projects

Neural Machine Translation System with baseline encoder-decoder architecture and attention mechanism.
Francisco Ramos
CS288: Natural Language Processing. University of California, Berkeley
Fine tuning LLMs to perform text-to-SQL task using transformers.
Francisco Ramos
CS288: Natural Language Processing. University of California, Berkeley

Personal projects

Pan & Tilt servo system autonomously controlled using Facial, Hand and Color Detection
Francisco Ramos Pérez, Carlos Álvarez Cia
December 2021
Code

Relevant Courses

• CS288: Natural Language Processing
EECS - UC Berkeley Fall '23
• CS198: Deep Learning for Visual Data
EECS - UC Berkeley Fall '23
• ME249: Machine Learning Tools for Modeling Energy Transport and Conversion Processes
ME - UC Berkeley Fall '23
• ME236: Control of Unmanned Aerial Vehicles
ME - UC Berkeley Fall '23
• ME292B: Artificial Intelligence for Autonomy
ME - UC Berkeley Spring '24
• Experiential Advanced Control Design
ME236 - UC Berkeley '24

Contact

You are very welcome to contact me regarding my research. I typically respond within a few days.
I can be contacted directly at francisco_ramos[at]berkeley.edu.