Guglielmo – Python, React, Next.js

The client's project was a "buy now, pay later" payment product designed to help businesses unlock cash flow by advancing 100% of the annual contract value (ACV) upfront while allowing customers to pay in flexible terms such as monthly or quarterly. It served as a sales enablement tool, offering flexible payment terms to customers and helping businesses close more deals. The product also included a revolving line of credit based on the company's ARR, typically capped at 60-70%.

Guglielmo's responsibilities included the following:

• worked closely with the founders and the product team to design a fast, responsive and clear web application that enabled vendors to close financed deals via Capchase;
• implemented the designs with modern UI/UX experience in React;
• created a complete Stripe integration that allows buyers to pay financed deals with tens of payment methods;
• handled the logic on a full-stack level, down to the business services and payment systems.

This project was a Django backend component for a young startup, positioning as a restaurant review aggregator that adheres to the Cozy Web principle: we aim to return to a more human-contact-based evaluation of restaurants and food venues. The idea was to create a system that allowed users to create reviews for restaurant and food venues, integrate with Google Maps and OSM to provide extensive places data and maps, and provide social interactions.

Among Guglielmo's responsibilities were the following:

• collected the requirements from the founders and designed and implemented a fully scalable backend architecture using Django;
• integrated with the Google Maps API and the OSM API in order to create the most fluid experience for users to search and collect restaurant data;
• wrote a full, pedantic suite of unit tests that run at CI time to ensure no bugs or regressions are introduced;
• deployed and managed all the infra for the scalable backend service, leveraging Docker containers.

This was a front-end part of the project for a young start-up—a restaurant aggregator. The idea was to build a fully cross-compatible web application that runs on browsers and Android. It would connect to its internal API and be fully unit-tested as well as strongly render-optimized.

Guglielmo's contributions were as such:

• cooperated with another freelance dev to create a cross-compatible and efficient web and mobile application;
• designed the architecture to be fully modular and inherently cross-compatible;
• had around 90% unit-test coverage, meaning all the components created were carefully tested;
• created automated CI/CD to deploy the application on the web and on the Google Play infrastructure.

The client's requirement was to perform automated KYB processes on buyers to assess their credit eligibility. Additionally, it was required to happen in a live fashion, with very low latency and low false negatives.

Guglielmo carried out the following tasks:

• integrated with third-party providers to perform risk assessments on financial entities;
• designed a smooth KYB experience in React for customers to be evaluated;
• created an iframe for big players in our customer base in order to integrate our KYB into their platforms;
• integrated with CRM software such as Hubspot and Salesforce to support KYB in those portals.

This company runs financial underwriting on hundreds of customers. The product is an in-house system integrating existing big data solutions, like DuckDBT and S3, with their own custom data processing pipelines written in Python. They delivered a working async solution that could handle hundreds of parallel processes, transform data, and build a solid metric system for our financial team to consume.

• Was directly involved with the design and implementation of the system from scratch;
• Designed a robust distributed system to execute DuckDBT pipelines in parallel;
• Connected the system with S3 instances, which allowed us to process massive amounts of data;
• Cooperated to design a testing environment that could cover all our e2e testing needs.

In order to properly consume the immense volume of data being streamed by their integrations, the team created ingestion pipelines that transformed our data lake of unordered data into a structured data warehouse, stored in GCP (BigQuery). The system had to be fast and most of all reproducible: they kept all the historical data in the lake at all times, and retransform everything on each day. Via BigQuery and Dagster pipelines with Python, they transformed tens of millions of JSON data in less than 8 minutes each day.

• Implemented multiple transformation pipelines in Python and DBT
• Created an automated system to compare our transformed data with metrics from the providers to ensure nothing was lost in the transformation
• Reached out to 3rd party providers to address inconsistencies on their API

During Gugliemo's 3 years in the company, they built over 15 web applications to empower the digital platform of a governmental institution (INAIL). These projects ranged from internal management software of welfare services to customer-facing portals to deal with sensitive insurance and retirement data. The customer facing projects were built using Angular and Java Spring, while their internal projects were mainly developed with React and Python.

• Worked on several full-stack architectures, from the UI implementation with close cooperation to product down to the database design and management.
• In the final year, directed a team of 3 people to implement a full-stack web portal in a timely manner, with strict static analysis and automated testing requirements.
• Integrated the architectures with private gov-owned networks.
• Created company-wide secure authentication systems with Firebase and Cloudflare.
• Handled production incidents promptly.