Maxime Bauer

I developed a legal chatbot application from scratch, combining a robust backend architecture, advanced AI integration (RAG), and a dynamic, high-end user interface.

Backend, Security & Data (FastAPI)

API Architecture: Full development of the FastAPI API, highly performant and typed with Pydantic for strict data validation.

Security: Integration of a secure OAuth 2.0 login system with session and user profile management.

Data: Cleaning, structuring, and formatting complex data streams.

AI Integration & Automation (RAG & LLM)

RAG Pipeline: Design of a semantic search system by querying a Milvus vector database to extract relevant legal context.

Prompt Engineering: Optimization of advanced prompts for the OpenAI API, ensuring accurate and reformatted legal responses.

WordPress Automation: Development of an automated bridge for one-click publishing of AI-generated legal content from the chat directly to a WordPress blog.

Frontend & UX/UI (React)

Dynamic Interface (React): Creation of a modern UI (nested components, smooth loading screens, theme management with Dark Mode).

UX Features: Implementation of Drag-and-Drop functionality and a system for organizing messages/folders in a nested structure.

Responsiveness: Optimization and full design of a fluid mobile version.

Stack: FastAPI, Python, Pydantic, React, OpenAI API, Milvus, OAuth 2.0, WordPress API.

For this pioneering client in the electric mobility sector in Morocco, I designed and developed the backend architecture and supervision interface for a central charging station management system (CSMS), managing IoT protocol integration and real-time communication.

IoT Architecture & Protocols (OCPP)

OCPP Implementation: Development of the central system (CSMS) based on the OCPP 1.6 protocol, ensuring secure bidirectional communication between charging stations and the cloud.

Asynchronous Processing: Design of an Asynchronous Worker dedicated to real-time OCPP message handling, ensuring high availability and decoupled processing of commands sent to stations.

Queue Management: Use of Redis to orchestrate the message queue and optimize worker performance.

Backend & Mobile Integration (FastAPI)

Mobile API: Architecture and development of the API layer (FastAPI) connecting the mobile application to the charging infrastructure.

Key Features: Integration of remote start/stop commands, real-time status tracking, and user transaction management.

Real-time Communication: Use of WebSockets for instant data feedback from the stations.

Supervision Interface & Dashboard (React)

Monitoring Dashboard: Complete development of a React dashboard for visualizing telemetry data and managing the overall fleet.

Administration: Integration of station and user management tools, real-time error reporting display, and direct database interaction tools.

Technical Stack: Python, Flask, WebSockets, OCPP 1.6, Redis, React.

For this assignment, I audited and migrated a Python/Flask application initially hosted on Google App Engine Flexible, whose infrastructure costs were disproportionate ($50/day excluding traffic), to a highly optimized and economical Serverless architecture.

Cost Optimization & Serverless Architecture

Cloud Run Migration: Replacement of App Engine with Google Cloud Run with minimal instance configuration down to zero, completely eliminating fixed costs when the application is not in use.

Containerization: Creation and configuration of an optimized Dockerfile to include Python dependencies and the FFmpeg system utility.

Handling Large Video Files (Up to 10 GB)

Signed URL Upload: Implementation of a signed URL system with Google Cloud Storage (GCS) to allow the Front-end to upload heavy files directly to the bucket, avoiding Flask server RAM overload.

Asynchronous Processing (FFmpeg): Integration of FFmpeg for on-the-fly video cutting and processing, with dynamic allocation of CPU/Memory resources only during the computation phase.

Frontend & User Experience

Loading Interface: Development of a dedicated user interface for uploading large files.

Real-time Tracking: Integration of a visual indicator to display the progress of video upload and processing for the user.

Technical Stack: Python, Flask, Google Cloud Run, Google Cloud Storage (GCS), Docker, FFmpeg, JavaScript.