Machine Learning Engineer

At Visagio, a company focusing on business development and transformation, Gustavo is the Tech lead of vBrain, an Artificial Intelligence as a Service startup that is spearheading the development of solutions for clients across various sectors. His main responsibilities involve guiding the team in designing and developing the platform, ensuring its seamless functionality and precise AI solutions. In this role he:

• Mentored junior and core developers, contributing to the growth and development of the team.
• Reduced the error rate for clients from 19% to 3% in a few months.
• Implemented a data lake structure and multiple airflow DAGs that reduced the execution time by 5x and optmized data storage.
• Improved inventory efficiency, reducing stock-to-sales ratio by 17%, and stockouts by 5%, and saving 12,000 labor hours through sales forecasting and stock optimization.
• Reduced 6 percentage points in default rate with a model for credit score definition solution in a B2B company

On previous projects:

• Led the end-to-end development of a Recommendations system, employing a rigorous approach involving testing and experimentation. The deliverables included hardware selection, robust data pipelines, AI Models, and a low-latency API for both the backend and real-time project KPI monitoring. The product he created revolutionized the user experience and delivered a 30% growth in revenue for its users in a food retail store.
• Guided the team in a growth and analytics project, skillfully translating complex data into actionable business insights and experimentation hypotheses. He presented findings to C-level executives, facilitating data-driven decision-making.
• Modeled a computer vision neural network to count the number of fruits that passed through a track.
• Orchestrated a series of experiments using A/B testing strategies to enhance the performance of a Retail company's website and email communications, optimizing user experience and doubling the conversion rate.

Lobo Guara was a fintech company dedicated to providing cutting-edge technology solutions for the investment advisory sector. Within 8 months of its inception, the company secured an impressive R$1 million valuation in angel funding.

Key offerings included:

• Developed a custom chatbot and dashboard, significantly enhancing inbound marketing capabilities and enriching profile data for client leads. This was implemented for two prominent brokerage firms with assets under management (AuM) totaling R$1.6 billion and R$200 million, respectively.
• Introduced an innovative Asset Analysis and Strategic Reallocation Proposal tool, resulting in a 30% revenue growth for a brokerage firm managing assets worth R$200 million.

• Developed engineering projects and proof of concept prototypes focusing on innovation, Industry 4.0, and IoT.

Winners of the innovation prize for the 2018 SpaceX Hyperloop Competition. https://tumhyperloop.de/

WARR Hyperloop Levitation pod

• Developed, prototyped, and tested multiple versions of rotating Halbach arrays for magnetic levitation and drive control.
• Designed, developed, and set up the LevPod and the test bench for the levitation module.
  1. Handled sensors, control, data acquisition, communication, and electrical design.
  2. Conducted Matlab simulations for the LevPod.

Enrolled in the PurECO Project and in the HR team.

Gustavo was involved in the DRONI project, which focused on developing a Robotic Airship with an Innovative Design. The project entailed creating an unmanned airship equipped with four electric vectored thrusters and aerodynamic surfaces as actuators, with a useful load capacity of 6 kg, primarily for environmental applications.

Project details:

• Hardware architecture: Designed a circuit board that facilitated the transition of the airship's control signal from manual to autonomous mode. It also enabled the embedded system to learn maneuvers by interpreting the manual control signal.

• Software architecture: Implemented a ROS (Robot Operating System) node to serve as a driver for the board, controlling the motors during autonomous operations. Developed a ROS node to monitor the airship's battery status.

• 2016 - Member of the team for the IEEE-VSSS Competition (Very Small Size Soccer), the Rover Project, and also was the lead Project Manager for the Robocode Team.

• 2015 - Helped in the development of a quadricopter and in the organization of Arduino, electronics, and programming Minicourses during the Control and Automation Week at Unicamp.

• 2014 - Participated in the Freescale Cup competition for the Track Follower Project.

• Automation projects management.
• Prospecting and negotiation of projects.