Benjamin G.

Design and development of an experimental software project combining artificial intelligence, automation, and data, aimed at analyzing, scoring, and finding movies and series available on streaming platforms based on geographic location.

🧱 Main Components

✷ LLM Finetuning

✷ Process Automation via n8n

✷ Containerized Architecture via Docker

✷ Cybersecurity at the core of project design

👨‍🔬 Technical Functioning of the Project (Summary)

⌘ Automated Architecture

Containerized architecture (Docker) orchestrated by n8n, enabling automation of the entire flow, from search to result delivery.

⌘ Intelligent Search and Structured Data

AI agents query the web in real-time to identify content availability on streaming platforms and automatically structure the results.

⌘ Specialized and Custom-Trained AI

Open-source language models (Mistral 7B, DeepSeek) are trained locally using LoRA fine-tuning techniques with MLX, from structured and enriched datasets. These models analyze results, assess audio/video quality, and select the most relevant platforms based on location, with reinforced training from user preferences (RLHF).

⌘ Automatic Location Management

The system automatically selects and activates the VPN connection country to access the platform chosen by the AI.

⌘ Simplified User Interaction

The user interacts via a Telegram bot (text or voice) to initiate a search and receives a confirmation indicating the selected platform and country.

⌘ Accessibility & Security

Controlled exposure of services via Cloudflare (DDoS protection, filtering, secure access) and implementation of cybersecurity by design practices (service isolation, secret management, network access limitation).

✷ Supporting multiple teams (3 to 5) within an agile at scale framework (SAFe)

✷ Facilitating Program Increment Plannings (PIP)

✷ Leading agile ceremonies and continuous improvement

✷ Implementing a Competency Center (CDC) to delegate the management of certain functional areas

Projects led:

⌘ Improving the agile maturity of certain teams (work on roadmap, Scrum board optimization, better delivery management): +25% velocity

⌘ Better planning and dependency management during PIs with Jira Plans

⌘ Supporting the transition of internal teams to a CDC (skill development, agile organization setup, continuous improvement...)

✷ Agile transformation of a team of 7 developers

✷ Steering a part of the internal platform and customer portal scope with 4 other Product Owners (agile at scale management)

Projects led:

⌘ Agile transformation of the development team (7 developers) - Scrum implementation: +25% velocity in 7 sprints

⌘ Creation of a real-time monitoring and prediction interface for a Japanese POC

⌘ Modification of contact management in the internal platform

⌘ Cross-functional management of the alerting redesign project > User Journey Mapping workshops with users > Consolidation of needs with User Story Mapping > Definition of the functional solution in collaboration with IT

How is a new feature developed at Energy Pool?

✔# User Journey Mapping workshops with users to understand their product usage

✔# Gathering pain points to identify issues and gaps

✔# User Story Mapping with a user, a developer, and the UX specialist to categorize and prioritize needs

✔# Definition of the functional solution in collaboration with IT