• At Fintron, he helped build part of the backend solution and gained his first professional experience working with Go. It was a challenging project with tight delivery deadlines and complex business rules, requiring him to quickly get up to speed while ensuring the solution remained reliable.
• He worked with Connected Performance during two separate engagements.
• In the first, as part of a consulting project, he helped identify several critical implementation issues that were causing application and service outages. He also contributed to defining a roadmap for improvements and participated in the initial implementation of those changes, helping the client gain greater confidence in the future of their product.
• In his current engagement with Connected Performance, he has been focused on investigating critical security and performance issues. His work aims to improve the product's reliability, scalability, and overall stability by analyzing complex technical problems and gaining a deep understanding of how different parts of the system interact.
• At Stamp, he had the opportunity to work across a wide range of areas as a software engineer. His responsibilities included building new backend services, developing high-volume data pipelines, implementing new frontend features, and contributing to infrastructure improvements. This experience allowed him to work across the full stack and develop a strong understanding of how all parts of the system fit together.

He was responsible for the creation, architecture, and development of various features for a SaaS platform called ScaleOps. This tool acted as an intelligent agent over a user-provided document base to facilitate information distribution within a company. Significant outcomes included reducing the learning curve for new employees and more effectively disseminating established standards and processes.

Proof of Concept - He independently initiated the code development of the platform using Python and other relevant technologies to validate the business model's viability.

Design System Development - In collaboration with a small multidisciplinary team, he contributed to deciding the architecture, tech stack, and implemented UI components using Typescript, React.js, and SASS for the library utilized in the application.

Use of LLMs - He was responsible for developing and maintaining the intelligent agent framework, leveraging Python for natural language processing and integration with large language models (LLMs).

Service Architecture and Deployment - He designed the initial schematic of the solution, including all services and security configurations on Amazon Web Services (AWS). The architecture was streamlined and secured, adhering to SOC2 recommendations.

He managed technical contracts, developed front-end solutions using React and back-end solutions with Node.js, creating seamless, scalable applications.

Additionally, he designed and implemented multi-brand design systems, ensuring consistency across platforms.

He also established quality standards and served as the go-to problem solver for complex technical challenges, optimizing performance and ensuring high-quality code delivery.

Adriano contributed to Design System consultancy projects, building Design System standards specifically for React applications and integrating them with Node.js backends.

He mentored new engineers on these standards, ensuring consistent development quality across both front-end and back-end solutions.

He led a project to build an intelligence engine for an application designed to help sales executives achieve their goals more easily and with minimal effort. With a team of just three people, he successfully operationalized this application for the entire company. The results were highlighted in the investor report for the quarter of its launch.

Solution Architecture - He was responsible for designing and implementing the application to ensure it was highly scalable and capable of making daily predictions for over 3,000 sales executives.

Data Ingestion and Analysis - He implemented an efficient data ingestion pipeline to process over 10 years of the company's sales history, providing the necessary data to feed the AI models.

AI Models - Using the necessary information, he built various predictive models that contributed to the daily recommendations provided to users. As a result, each sales executive received a personalized recommendation tailored to their profile. This approach also helped in developing training strategies for groups with performance expected to be below average.

He worked on an AI initiative within the product department, where he developed various models to help the platform categorize, select, approve, and recommend candidates from various fields. These models were instrumental in matching candidates with companies that had specific hiring needs for professionals.

Machine Learning Models - He developed clustering models, resume analysis and scoring models, and ranking and sorting models. These models were utilized in various workflows and processes across the platform.

Dashboards and Reports - Leveraging the extensive data from the platform, he developed various dashboards and reports for the product and growth teams to monitor and improve the features in production.

Monitoring - He designed and implemented monitoring mechanisms to detect performance drops in production models, primarily caused by market changes, seasonality, and alterations in business rules. These mechanisms simplified bug identification and helped maintain the models effectively.

The goal was to implement a model that combined intelligence with custom-built tools, aiming to reduce repetitive work and improve information management.

From Concept to Delivery - He worked on projects from conception to implementation as a full-stack engineer, aligning with client tech stack requirements and ensuring proper documentation for future maintenance.

AI Application Development - He developed artificial intelligence applications such as recommendation systems, ranking systems, and Natural Language Processing (NLP) systems to meet operational requirements.

Business Rules - He gained significant insights into the diverse needs of the clients served, which enhanced his understanding of the dynamics of the Brazilian market.