I focus on building deterministic systems in a non-deterministic world. I lean toward backend complexity, AI/ML integration, and simple, effective frontend delivery.

Upload PDFs and have intelligent conversations with your documents — powered by AWS Bedrock, deployed on EC2, with session-isolated S3 storage.

Tech Stack React + Vite | FastAPI | AWS Bedrock (Llama 4 Scout + Titan Embeddings V2) | ChromaDB | Amazon S3 | Docker | AWS EC2

Engineering Highlights

• Migrated from local Ollama setup to AWS Bedrock — swapping nomic-embed-text for Titan Embeddings V2 and Llama 3.2 for Meta Llama 4 Scout 17B.
• Session-based document isolation ensures each user only queries their own uploaded PDFs.
• Async background processing pipeline with status polling — frontend waits for full embedding before firing initial query.
• S3 integration for persistent, scalable PDF storage replacing local Docker volumes.
• Rate limiting via slowapi (10 req/min chat, 5 req/min upload) with IAM role-based auth — no hardcoded credentials anywhere.
• OCR fallback via pytesseract ensures scanned pages are never skipped.
• Real-time streaming responses via Bedrock's converse_stream API.

🌿 Local version (Ollama, no cloud required) available on the main branch.

https://github.com/VeeSeven/RapidSummarize

Enables complex note hierarchies (folders in folders / notes in notes) with an intuitive tagging system that auto-assigns consistent colors for visual organization.

Tech Stack React + Vite | Django REST Framework | PostgreSQL | JWT Authentication | Docker

Engineering Highlights

• Fully containerized with Docker Compose for one-command local development.
• Implements recursive note relationships and efficient tag filtering via Django ORM.
• Rich text editor (TipTap) with auto-save, word count, and formatting toolbar.
• JWT authentication with token rotation and blacklisting for enhanced security.
• Responsive UI built with Chakra UI and state management using React hooks.

https://github.com/VeeSeven/MindSpace

class Developer:
    def __init__(self):
        self.languages = ["Python", "JavaScript", "Java", "SQL"]
        self.backend   = ["FastAPI", "Django", "Flask", "REST APIs"]
        self.frontend  = ["React", "Tailwind", "ChakraUI"]
        self.ai_ml     = ["RAG", "LLM Integration", "Vector Search", "AWS Bedrock"]
        self.infra     = ["AWS EC2", "AWS S3", "AWS IAM", "Docker", "PostgreSQL", "ChromaDB"]