About Me
I’m a Software Engineer specialized in full stack development and AI. I design, build, and deploy scalable web and mobile applications, develop AI-powered solutions, and optimize machine learning models for real-world performance. My expertise spans end-to-end system architecture, rapid prototyping, API development, cloud deployment, and implementing AI workflows from experimentation to production.
My Expertise
Technologies and tools I work with to build modern, scalable applications
Python
JavaScript
TypeScript
SQL
C# / .NET
C
React
Next.js
Tailwind
Framer Motion
HTML/CSS
Bootstrap
Responsive & Mobile-First Design
UI/UX Optimization
Node.js
Express.js
AWS Lambda
AWS API Gateway
AWS Amplify
AWS CloudFront
AWS Cognito
AWS IAM
AWS CloudFormation
AWS Elastic Beanstalk
AWS Key Management Service (KMS)
Vercel
Serverless
REST APIs
GraphQL APIs
Authentication & Security (JWT, OAuth)
Python
JavaScript
TypeScript
SQL
C# / .NET
C
React
Next.js
Tailwind
Framer Motion
HTML/CSS
Bootstrap
Responsive & Mobile-First Design
UI/UX Optimization
Node.js
Express.js
AWS Lambda
AWS API Gateway
AWS Amplify
AWS CloudFront
AWS Cognito
AWS IAM
AWS CloudFormation
AWS Elastic Beanstalk
AWS Key Management Service (KMS)
Vercel
Serverless
REST APIs
GraphQL APIs
Authentication & Security (JWT, OAuth)
RAG pipelines
LLM prompt engineering
Google Gemini
ChatGPT
Claude-Sonnet
Amazon Q
scikit-learn
PostgreSQL
telemetry & monitoring
GitHub Actions
Supabase
Firebase
Firebase Authentication
Firebase Firestore
Firebase Realtime Database
Firebase Hosting
Firebase Cloud Functions
Firebase Cloud Messaging (FCM)
CloudWatch
CI/CD
Docker
Kubernetes
Serverless Framework
Product → deploy ownership
rapid prototyping
bilingual (PT/EN)
Spanish
English
Portuguese
RAG pipelines
LLM prompt engineering
Google Gemini
ChatGPT
Claude-Sonnet
Amazon Q
scikit-learn
PostgreSQL
telemetry & monitoring
GitHub Actions
Supabase
Firebase
Firebase Authentication
Firebase Firestore
Firebase Realtime Database
Firebase Hosting
Firebase Cloud Functions
Firebase Cloud Messaging (FCM)
CloudWatch
CI/CD
Docker
Kubernetes
Serverless Framework
Product → deploy ownership
rapid prototyping
bilingual (PT/EN)
Spanish
English
Portuguese
Experience & Projects
AI software developer
Revelo
Remote • Jan 2026 – Present
Selected for invite-only AI Software Development role focused on improving LLM performance in real-world software development tasks through advanced evaluation, red-teaming, and systematic failure analysis.
Key Responsibilities:
- Analyzed and corrected failure modes in 300+ complex code generations across Python, JavaScript, and TypeScript, increasing execution correctness and reasoning reliability by 25%.
- Designed structured evaluation criteria and feedback loops that reduced logical errors and invalid assumptions in multi-step model outputs, improving downstream task success rates.
- Implemented systematic red-teaming strategies to identify edge cases, boundary conditions, and potential failure scenarios in LLM-generated software solutions.
- Collaborated with ML research teams to refine model training pipelines, contributing domain-specific insights that enhanced code generation quality and architectural soundness.
- Developed comprehensive documentation of common LLM pitfalls and anti-patterns in software engineering contexts, establishing best practices for prompt engineering and output validation.
- Conducted comparative analysis of model performance across different code complexity levels, providing data-driven recommendations for model selection and optimization strategies.
Machine Learning Specialist
Outlier
Remote • Jun 2025 – Dec 2025
Specialized in LLM evaluation, debugging AI-generated code, and enhancing model outputs through systematic analysis and structured feedback mechanisms.
Key Responsibilities:
- Debugged and evaluated 500+ LLM-generated Python and JavaScript code samples, improving executable pass rates by 20–25% through targeted error identification and correction.
- Identified recurring reasoning and architectural failure patterns in AI-generated solutions, including off-by-one errors, improper error handling, and inefficient algorithms, reducing error frequency by 15%.
- Improved prompt and task structure to minimize hallucinations and increase determinism in technical outputs, establishing clearer constraints and validation criteria.
- Conducted in-depth analysis of model reasoning chains to identify logical inconsistencies, invalid assumptions, and gaps in contextual understanding.
- Developed and maintained comprehensive evaluation rubrics for assessing code quality, correctness, efficiency, and adherence to best practices across multiple programming paradigms.
- Collaborated cross-functionally with AI trainers and prompt engineers to iterate on task design, resulting in more robust and reliable model performance.
- Documented patterns of model failure modes and success cases, contributing to knowledge base that informed future training data selection and model improvement initiatives.
Full Stack Engineer
StudyShield
2025
Architected and implemented StudyShield, an AI-powered PWA using React 18, TypeScript and Vite; delivered a responsive, installable web app with offline support and service-worker versioning.
www.studyshield.site
Key Responsibilities:
- Designed and shipped a dual-mode AI platform integrating Google Gemini (Flash online + Nano offline) with streaming responses and automatic failover to ensure low latency and offline resilience.
- Integrated Supabase for Auth, PostgreSQL and Storage; authored DB migrations and Row-Level Security policies and designed user_profiles and conversations schemas to enforce per-user data isolation.
- Built secure multimodal chat: file attachments (images/PDFs via pdfjs), voice input (Web Speech API), content sanitization (DOMPurify) and AI content filtering to protect users.
- Implemented Focus Mode and distraction-blocking UX to increase study session productivity; added gamified progress tracking and analytics dashboards (Recharts) to boost engagement.
www.studyshield.site
Technologies Used:
Full Stack Developer
Megaphoton – Solar Energy
Remote • 2025
Bilingual conversational assistant powered by Google Gemini 2.5-Flash with RAG pipeline and WhatsApp escalation for solar energy company.
www.megaphoton.com.br
Key Responsibilities:
- Implemented a bilingual (PT/EN) conversational assistant powered by Google Gemini 2.5-Flash, with language detection and persistent context.
- Built a RAG pipeline + knowledge base and verification layers to minimize hallucinations and ensure company-data correctness.
- Implemented WhatsApp escalation for human handoff and quick-action flows for common requests; improved customer self-service by 25%.
- Production deployment: Vercel serverless functions, env-secure API integration, rate-limit planning, robust fallback/error handling, mobile-optimized UX (React + TypeScript + Framer Motion + Tailwind).
www.megaphoton.com.br
Technologies Used:
Full Stack Developer
MediMentor
Remote • 2024 – 2025
AI-driven medical platform providing symptom analysis and differential diagnosis with clinical-grade data sourcing.
main.do8zwgfpt20yc.amplifyapp.com
Key Responsibilities:
- Integrated Isabel Healthcare API to power AI-driven symptom analysis and differential diagnosis, ensuring clinical-grade data sourcing.
- Architected serverless backend (AWS Lambda, API Gateway) with secure auth (Cognito) and production monitoring (CloudWatch), enabling reliable real-time interactions.
- Implemented frontend hosting and CI/CD via AWS Amplify and CloudFront for fast, globally distributed delivery.
- Built analytics dashboards and monitoring to track system health and performance (telemetry & logging).
main.do8zwgfpt20yc.amplifyapp.com
Technologies Used:
Full Stack Developer
Muscles & Balance
Remote • 2023 – 2025
Full-stack health & fitness platform with secure authentication, workout tracking, and interactive analytics dashboards.
muscles-and-balance-7.web.app
Key Responsibilities:
- Built first full-stack health & fitness platform with secure authentication, workout tracking, and interactive analytics dashboards; established CI/CD pipelines (AWS Amplify, GitHub Actions) for rapid, reliable releases.
- Implemented realtime UX features (progress tracking and dashboards) and telemetry to measure performance and user flows.
- Led end-to-end development of a fitness app: product design → auth → real-time dashboards → CI/CD and deployments. This project passed by MLH Fellowship which validates its maturity.
muscles-and-balance-7.web.app
Technologies Used:
Get in Touch
Have a project in mind, a question, or just want to connect? Feel free to reach out!
Send Me a Message
I'll do my best to respond as soon as possible.
Contact Information
Alternatively, you can reach me through the following channels:
Preferred Contact Method:
Email is generally the quickest way to get a response for inquiries. For professional networking, LinkedIn is also a great option.