Marcio Oliveira
AI-Driven Full Stack Developer
Senior AI & Full Stack Engineer with a decade of experience building production-grade AI systems, real-time data pipelines, and intelligent automation workflows. Expert in Python, deep learning, and scalable backend infrastructure. I specialize in integrating LLMs and AI models into robust platforms, optimizing performance, and engineering seamless user experiences using React, Next.js, and modern Python frameworks like FastAPI, Django, and Flask. Whether it's cloud-native AI deployment or end-to-end app delivery—I'm focused on making innovation real.
View My Work
Technical Skills
AI & Machine Learning
- Machine Learning
- Deep Learning
- NLP (spaCy, OpenAI GPT)
- Computer Vision (OpenCV)
- Reinforcement Learning
- LangChain
- LangGraph
- LlamaIndex
Backend & APIs
- Python
- FastAPI
- Flask
- Django
- Node.js
- RESTful APIs
- GraphQL
- Jakarta EE
Frontend & Web
- React
- Next.js
- HTML5
- CSS3
- TailwindCSS
Databases & Vector Stores
- PostgreSQL
- MySQL
- MongoDB
- Oracle (PL/SQL)
- Neo4j
- Pinecone
- FAISS
ML/DL Frameworks & Libraries
- Scikit-learn
- TensorFlow
- PyTorch
- Keras
- XGBoost
- U-Net
- DeepLabV3+
Cloud, Tools & Automation
- AWS (S3, Lambda)
- Firebase
- Render
- Railway
- Docker
- Git
- n8n
Most Successful Projects
AI Powered Call Center Solution
Led the full-stack development of a real-time AI-powered call center platform using Next.js (frontend), Django (admin & auth), and FastAPI (AI services). Architected and implemented core features including real-time transcription with FasterWhisper, intelligent query handling via Google LLM, semantic search with Pinecone vector DB, and RAG-based pipelines for accurate AI responses. Developed a smart escalation mechanism for unresolved queries and enabled live WebSocket communication. The system supports multilingual interaction (English and Arabic) and includes a modular admin dashboard for service management Builted in a month. MVP stage.
ChatLLM Teams
Architected and implemented the backend for a multi-user enterprise conversational assistant platform leveraging Abacus.AI. Focused on real-time LLM orchestration, I developed token-based access control, dynamic memory context handling, and multi-agent prompt chaining logic. The platform supports secure, scalable collaboration across teams, enabling contextual AI-driven conversations, agent coordination, and persistent user sessions. My work emphasized reliability, latency optimization, and modular design for enterprise deployment. Built in 7 month, Real Production
Real Estate Property Finder
Independently developed a full-stack, space-themed real estate discovery platform using Next.js for the frontend and Django for the backend and admin panel. I implemented a custom vector search engine for intelligent property matching based on semantic similarity, integrated geolocation filtering, and built LLM-powered modules for dynamic property description generation. The platform also features a responsive UI, user authentication, and an optimized deployment pipeline designed for performance and scalability in production environments. Built in 4 month, Real Production