Luciana Cendon

Engineer turned first-time founder, building AI-native systems

About me

I’m Luciana, an engineer and researcher now building my first company. I bring over 9 years of experience in machine learning, computer vision, and data science. I hold two master’s degrees: one in Electrical Engineering with a focus on Computer Science from Caltech, and another earned through a collaborative program across EPFL, Grenoble INP, and Politecnico di Torino, which provided me with a strong foundation in engineering and advanced technologies.

Over the years, I’ve contributed to impactful projects, from building real-time full-stack computer vision pipelines for robotics applications at Quiet Machines to advancing machine learning and data systems at NASA’s Jet Propulsion Laboratory. These experiences sharpened my expertise in end-to-end ML pipelines, data analysis, and applied AI, while deepening my understanding of real-world data-driven problem-solving.

Since 2024 I've been in founder mode, exploring multiple AI product directions through prototyping and user testing. I've built prototypes across matchmaking systems, workflow automation, and recommendation engines, gathering feedback from beta users and learning what it takes to validate ideas and iterate on product direction. I'm currently focused on agentic AI systems and structured reasoning applications, working on system architecture and evaluation frameworks for multi-step autonomous workflows.

Alongside my AI work, I’m starting a small AP STEM program focused on preparing students for AP Physics and AP Calculus exams. More info here.

When I’m not immersed in tech projects, I enjoy honing my musical skills on the electric guitar 🎸 and staying active through exercise and hiking. I’m also a coffee enthusiast with a passion for crafting the perfect espresso ☕. Most importantly, I’m a proud mom to an amazing daughter, and we love traveling 🌍 together and discovering new places.

Education

California Institute of Technology (Caltech)

Master of Science in Electrical Engineering

Pasadena, CA - June, 2016

École Polytechnique Fédérale de Lausanne (EPFL)

Grenoble Institute of Technology (Grenoble INP)

Politecnico di Torino

Master in Micro and Nanotechnologies for Integrated Systems

Conjoint degree from three universities

Lausanne, Switzerland • Grenoble, France • Turin, Italy - March, 2014

Politecnico di Torino

Bachelor in Electronics Engineering

Turin, Italy - March, 2012

Professional Experience (Recent)

Founder & Technical Lead

Stealth AI Startup

June 202 - Present | Los Angeles, CA

Built and tested multiple AI product prototypes across matchmaking systems, workflow automation, and recommendation engines using diverse tech stacks. Conducted user validation with beta testers; iterated on product direction based on technical feasibility and market feedback. Currently focusing on agentic AI systems and structured reasoning applications

Senior Computer Vision/Robotics Engineer

Quiet Machines, LLC

Nov 2022 - May 2024 | Pasadena, CA

I was the sole computer vision engineer responsible for designing and implementing the entire computer vision pipeline. My focus was on object detection and tracking within video contexts, particularly isolating runners in complex, noisy environments. I developed and trained models for image segmentation and scene understanding, managing data curation and annotation to improve detection accuracy. Additionally, I optimized and integrated these models into the prototype system for high-performance inference and conducted field tests to enhance system performance. I also developed and integrated the frontend with the backend, ensuring seamless communication and real-time data synchronization via WebSockets.

Professional Experience

Data Scientist & Software Engineer

Jet Propulsion Laboratory

Nov 2019 - Nov 2022 | Pasadena, CA

At JPL, I contributed to multiple projects, including building machine learning pipelines and developing data analysis systems. I supported NASA’s ARIA Project by designing anomaly detection methods using spatio-temporal time series data to predict volcanic activity. For NASA's Ocean Worlds Life Surveyor (OWLS) Project, I created high-performance simulation tools for generating synthetic holographic images and developed algorithms for automated peak detection in noisy spectrometer data. I've also built automated dashboards in Power BI to visualize resource utilization, supporting data-driven decisions.

Research Engineer

California Institute of Technology

Jan 2019 - Oct 2019 | Pasadena, CA

I spent 9 months working at Pietro Perona's lab, where I developed, trained, and deployed detection and pose estimation models using AWS SageMaker. These models were used as part of a system for automated behavior analysis in lab animals, facilitating scalable cloud-based processing for pose estimation and detection. During this time I also designed a system for automated video editing using imitation learning. This project is available in the projects section.

Research Engineer

Oben Artificial Intelligence

Dec 2017 – Jan 2019 | Pasadena, CA

At Oben, I developed advanced software solutions for personal avatar animation, including implementing gaze tracking using C++ and ensemble regression trees. I also worked on extracting lighting information from facial textures using augmented spherical harmonics, as well as developing an active appearance model for automated lip synchronization. Additionally, I designed a general-purpose chat system adaptable to new personalities and delivered Q&A chat software, which served as the foundation for the company's mobile health initiative.

Research Engineer

HRL Laboratories

Oct 2016 – Nov 2017 | Pasadena, CA

At HRL, I worked on several projects including predictive modeling for well optimization, leveraging data analysis to improve efficiency. I also contributed to the development of reinforcement learning algorithms for autonomous airplane flight and worked on low-power image recognition systems.

Research Engineer

California Institute of Technology

June 2016 – Sept 2016 | Pasadena, CA

During my time at Caltech, I conducted data mining for mouse behavior analysis in time-series data, utilizing Markov chains to model and analyze complex behavior patterns. This work helped identify important behavioral trends, contributing to ongoing research efforts in neuroscience.

Blog

A collection of my writing inspired by my curiosity and passion for learning. This space is where I share knowledge and document my journey through the mathematics of machine learning, cutting-edge AI models, the real-world impact of AI, and beyond.


  • All
  • Technical
  • Non-Technical

How Diffusion Models Work: A Beginner's Guide to DDPMs

This article simplifies and explains the math-heavy foundations of diffusion models, focusing on core processes like the forward and reverse diffusion steps, Markov chains, and Gaussian distributions. By breaking down the principles that underpin DDPMs, it makes the underlying mathematics accessible while maintaining clarity and focus on foundational understanding.


Introduction to Diffusion Models

An overview of diffusion models, their significance in generative AI, and how they compare to traditional models like GANs and VAEs. This article explores their core principles, key advantages, and real-world applications, such as text-to-image generation, inpainting and beyond.


Smooth Imitation Learning

This project highlights the application of the Smooth Imitation Learning (Simile) algorithm, developed by Hoang M. Le and colleagues, to the problem of automated video editing, utilizing its sequence prediction capabilities. It offers a detailed explanation of the algorithm, presents experimental results, and includes the custom code developed for its implementation.


How SMBs Can Harness AI: A Guide to Efficiency and ROI

Discusses how small and medium-sized businesses (SMBs) can leverage artificial intelligence to enhance efficiency, boost return on investment, and maintain competitiveness. It provides actionable insights and examples to guide SMBs in implementing AI strategies effectively.

Contact

Let's Connect!

I’m always open to connecting with fellow founders, investors, and collaborators who are building the future of AI and technology. If you’d like to talk about my work at Marvus, potential collaborations, or just swap ideas, reach out below.


Email

luciana.hpcendon@gmail.com

Linkedin