As the sole technical lead, I developed a SaaS platform that uses AI to enhance meeting productivity by transcribing, summarizing, and extracting key insights from Google Meet and Microsoft Teams sessions. I managed full-stack development, DevOps, and AI integration, providing businesses with actionable insights and improved visibility into their meetings.Technologies & Achievements:- Cloud Infrastructure & Deployment: Designed and managed the entire infrastructure on Google Cloud, implementing production and UAT environments for smooth testing and deployment. Built automated CI/CD pipelines using GitHub Actions, streamlining code deployment across environments. Managed Google Cloud IAM for secure, efficient access control without unnecessary complexity.- Backend Development: Built a robust serverless backend using Firebase Functions (over 100 functions), triggered by events such as database writes and HTTP requests. Integrated Firestore for real-time data sync, and synchronized it with a Neo4j database for advanced AI processing. Managed critical services for meeting audio transcription, diarization, and AI-powered summarization using WhisperX (Python), running in Google Cloud Batch with Docker containers. Automated AI meeting attendance with services leveraging Puppeteer, and implemented Google Cloud Pub/Sub for event-driven architecture.- Frontend: Developed a responsive, multilingual dashboard using Next.js 14, React, TypeScript, Tailwind CSS, and Shadcn UI components, leveraging server-side rendering for performance optimization. Implemented Firebase Auth for secure user authentication, including email/password and social logins, as well as multi-factor authentication (MFA). Focused on real-time data synchronization and user experience while building a workspace feature for hierarchical company structures and Stripe subscription management.

Responsible for developing automated investment strategies, optimizing both returns and losses on a medium to long term financial horizon. To do so, I managed to: - Architect, design and develop a modular, dynamic and easy-to-use framework that allows to, systematically, create and test new investment strategies, and backtest them to, effortlessly, find the best strategy approprieted to the metrics at hand;- Implement many processing filters that follows both technical and fundatamental analysis, aswell as many portfolio optimization methods;- Develop a robust backtest plataform and a simulation feature to compare tens of thousands of different automated strategies;- Successfully implement 5 distinct automated strategies in a production environment to be consumed by the company costumers.

Assisted in the developtment of many machine learning models to forecast electrical energy pricing in Brazil to costumers like JBS S.A and Eldorado Brasil. In doing so a managed to: - Automate, optimize and produce machine learning models based on decision trees and deep neural networks applied to Brazil's electrical sector; -Implement and automate data pipelines to serve climatological information from climate forecasting models (ECMWF and CFSv2) to our models in production.

Studied both data analysis and machine learning modeling techniques applied to biological signals, like EMG and EEG. This was my first contact with machine and deep learning approaches to modeling problems.