Juan Lozano

Co-founder & CTO

Mariana Minerals

Sep 2024 - Current (1 year 11 months)

Member

Inception Studio

Apr 2023 - Current (3 years 4 months)

United States

S

Senior Staff Machine Learning Engineer

Mar 2023 - Aug 2024 (1 year 6 months)

Exploring and ML consulting

Senior Staff Machine Learning Engineer

Selfemployed

Mar 2023 - Aug 2024 (1 year 6 months)

Exploring and ML consulting

Senior Staff Machine Learning Engineer

Affirm

Jun 2021 - Feb 2023 (1 year 9 months)

Affirm acquired Kite. I primarily worked on the GrowthML team where I architected the first multi-armed bandit system to optimize marketing spend by targeting users that would have positive ROI for Affirm after loan take up. I also co-architected Affirm's first ad system to target users on the Affirm app with relevant ads that maximized ROI. Additionally, I built Affirm's first system to use ML to proactively reach out to users that were at risk of churning in order to encourage them to continue using our product. Our team was later transitioned to the Credit Underwriting org where I worked on Affirm's ML Underwriting models.

Head of Machine Learning

Kite

Apr 2017 - Jun 2021 (4 years 3 months)

Built ML tech and team from 0-1 and scaled the user base to half a million monthly active developers. I worked directly with the rest of the leadership team to research, plan, execute, and deliver novel product experiences. I pitched the technical vision for Kite to series A investors, helped pitch the engineering team to potential acquirers, and advocated for, and oversaw, our transition from classical ML to deep learning. My team specifically focused on combining machine learning and program analysis techniques to learn high-level semantic representations of source code to enhance developer productivity through various product experiences (autocomplete, bug detection and correction, documentation generation, and code search). I built

Machine Learning Engineer

Kitecom

Aug 2015 - Apr 2017 (1 year 9 months)

Kite is an AI assisted programming copilot. My work focused on combining machine learning and program analysis techniques to learn high-level semantic representations of source code to enhance developer productivity through a variety of product experiences (autocomplete, bug detection and correction, documentation generation, and code search).

I

Founding Engineer

Insidemaps

Jul 2013 - Aug 2015 (2 years 2 months)

Wrote many of the core 3d reconstruction algorithms, went through StartX with the two founders, and helped raise multiple seed rounds. Primarily focused on writing the core 3d reconstruction algorithms including: - Semi-definite programming techniques for robust estimation of camera positions and orientations. - Large scale bundle adjustment incorporating points, lines, camera positions, and orientations. - Randomized correspondence techniques for enhancing the density of sparse point clouds. - Graph cut approaches for extracting triangular meshes from dense point clouds. - Variational mesh refinement based on gradient descent using a multi-resolution adaptive mesh. - Variational mesh segmentation based on learned features. - Object detect

B

Cofounder and CTO

Blusofa

Mar 2012 - May 2013 (1 year 3 months)

Stealth startup, more to come soon.

Software Developer

Sony Mobile Communications

Mar 2011 - Mar 2012 (1 year 1 month)

Member of a five person team designed to operate semi-autonomously within Sony Mobile, in order to rapidly develop and pitch novel mobile applications to engineering and marketing teams. One of the main projects for the summer was the development of a beta version of a location based Android mobile application. At the end of the summer this project was taken over by a full engineering team that is currently preparing the application to be distributed on all 10 million of Sony Mobile's Android phones.

Software Developer

Adobe

Jun 2010 - Sep 2010 (4 months)

Design and implementation of a test environment for Adobe project teams to deploy and test new library components for the project in an isolated, automated, environment. The test environment allows developers to manipulate the display properties of the new components and see the results in real time. Additionally, developers have the ability to feed test data scripts into the component to automate functionality tests. Finally, the application provided example use cases for the library features in order that other members of the team could quickly learn how to use the new components.

About

Detail

Timeline

Résumé