Juan – AWS, React, Python

An AI-powered web app to accelerate the learning process of university students. It helps students to write theses and articles related to different topics.

• Implemented the entire back-end and front-end logic (excluding visuals) using Next.js;
• Developed all AI features integrating the Gemini API (Google's LLMs);
• Participated in the UI/UX discussions;
• Presented multiple ideas to the client, some of which were later transformed into features;
• Managed infrastructure (Vercel + GCP) and version control (Github).

An AI-powered news aggregator that leverages vector embeddings and LLMs to semantically relate and consolidate news from multiple sources, providing a comprehensive and pluralistic view of events.

• Completely built the AI engine that powers the platform using Python, Shell scripts, Local Ollama, Flask, etc;
• Actively participated in the FE development, mainly in the FE logic (Next.js, React);
• Set up the infra in AWS and Vercel, Docker images, and the CI/CD logic by using Github actions;
• Proposed the idea and worked on it before getting down to work.

An R&D project demonstrating the power of early LLMs (GPT-3) in supporting the legal department's internal tasks. It enhanced the review process for various legal documents, highlighting critical issues (red-flagging) and helping the legal team save significant time.

• Presented the idea;
• Collaborated with the legal team to define the necessary features;
• Implement the prototype.

A fintech startup that develops a platform providing financing for exporting SMBs in LATAM. It is revolutionizing access to international trade, with a team of experts working tirelessly to provide the best tools and financing solutions for trading companies without access.

• Designed the back-end architecture;
• Implemented a significant portion of the back-end foundations;
• Co-designed the CI/CD pipeline with the DevOps team;
• Integrated various financial data providers;
• Developed the unit test suite.

An omnichannel contact center that provides easy-to-use communication capabilities (telephony, WhatsApp, email, etc.) for many companies across LATAM.

• Implemented the core backend and many features of the web platform;
• Maintained legacy code (PHP, Python 2.x);
• Discussed and planned feature implementations;
• Managed versioning of the platform (Jenkins scripts, git support, etc);
• Assisted with front-end development when the team was overwhelmed.

Helixia.io is a Chatbase AI solution that automates sales inquiries and support workflows for manufacturing businesses, enabling real-time quote generation, lead qualification and 24/7 customer service without expanding human teams.

• Built the backend from the ground up
• Designed and implemented AWS infrastructure and CI/CD pipelines from scratch
• Collaborated closely with frontend developers and contributed to task planning activities

For InDebted.co, I implemented AI-driven translation scripts to rapidly and massively localize their website in over five countries, integrating Hugo’s native i18n switching, LLM-driven batch translation with confidence-based review, RTL layout adjustments, dynamic regional currency formatting, and embedded compliance phrasing.

• set up project and local execution environment, including essential fixes
• scripted enumeration of Markdown and JSON assets for localization using regex filtering
• extracted hard-coded strings from templates into Hugo’s i18n framework for standardized translations
• reviewed redirects and routing logic to ensure accurate language-specific URL delivery
• built AI-driven translation scripts to translate and optimize all content