Guilherme – Python, LLM, AWS

A Transportation Management System (TMS) platform for managing fleet information and providing insights that help carriers improve operational efficiency.

• US Transportation Management System (TMS): Serving as Full-Stack Engineer for a Chicago-based logistics client, driving the end-to-end development of their core enterprise platform.
• Government API Integration & Compliance: Architected an inspection control system with direct API integration to the USDOT (U.S. Department of Transportation), automating infraction tracking and ensuring strict regulatory compliance.
• Driver Ranking Algorithm: Develop from scratch an algorithm to measure in real time how safe a driver is based on GPS data, speed, sudden movements, shift duration, and other factors.

A software development consulting engagement for a Chicago-based logistics company, focused on delivering software solutions that support its business operations and logistics processes.

• End-to-End SaaS Development (Python/React): Architected and deployed 5+ greenfield web applications from scratch (Zero to One). Built responsive user interfaces using React.js, Tailwind CSS, and ShadCN/UI, integrated with Axios for seamless REST API communication.
• Industrial Optimization System (MES): Developed a secure, multi-tenant Manufacturing Execution System backend using Django. Implemented complex production scheduling algorithms using Python, Pandas, NumPy, and Google OR-Tools to solve Linear Programming problems, exposed via REST APIs.
• Secure AI Assistant Integration: Architected a conversational AI assistant within the MES using Gemini. Utilized Function Calling to allow the model to query real-time company data, enforcing strict tenant isolation and information security protocols through Role-Based Access Control (RBAC) to ensure zero data bleed.
• Enterprise Document Understanding (AI/RAG): Built systems for complex document understanding, constructing an OCR and RAG architecture. Orchestrated high-volume asynchronous workflows with Redis and BullMQ between extraction services and LLM providers.

A B2B2C marketplace serving both customers and sellers, with dedicated product areas focused on improving experiences and supporting interactions across both sides of the marketplace.

• Technical Leadership & Architecture: Bridged the gap between product and engineering by conducting code reviews, designing system architecture (Database schemas, API contracts), and ensuring technical feasibility of roadmap items.
• LLM & Data Prototyping: Hands-on development of Python scripts and PoCs. Created an automated pipeline using LLMs to classify high volumes of user support tickets, directly feeding insights into strategic roadmap planning.
• Algorithm Optimization: Led the revitalization of the psychologist search algorithm using data analysis, resulting in a 20-point NPS increase and reducing churn by 90% in 6 months.

An online psychotherapy health-tech platform connecting users with mental health professionals and supporting access to remote therapy services.

• High-Performance Python APIs: Designed and maintained scalable RESTful APIs using Django Rest Framework (DRF) to serve a decoupled React SPA (Single Page Application), ensuring strict contract adherence and low latency.
• Data-Intensive Ranking System: Engineered a custom ranking algorithm using NumPy and Pandas for matrix calculations. Optimized performance by offloading heavy math computations to asynchronous tasks via Celery/Redis, ensuring pre-calculated data availability for instant database retrieval.
• Infrastructure & Microservices: Deployed NLP-driven microservices for recommendations using Docker on AWS. Managed a hybrid database architecture (MariaDB for relational data, DynamoDB for high-velocity logs).
• Environment: Python 3, Django, Celery, Flask, Docker, AWS (EC2, DynamoDB), MariaDB, Spacy, BERT.