Ali Darouy

Partnering with SMEs (PME) and Mid-market companies (ETI) to bridge the gap between their business data and AI-driven growth. I specialize in building custom, production-ready AI tools and automated data systems that solve operational inefficiencies and unlock new revenue streams. Key Services & Impact: AI Strategy & Implementation: Helping business leaders identify high-impact AI use cases and executing them from initial prototype to full production deployment. Custom AI Tooling: Developing bespoke AI applications (LLM integration, automated document processing, intelligent agents) tailored to specific industry workflows. Robust Data Integration: Leveraging my background in high-scale IaaS to build reliable, type-safe data pipelines that ensure AI models operate on high-quality, real-time business data. Product-Led Engineering: Managing the end-to-end product lifecycle—from business requirements and technical architecture to UI/UX and cloud deployment. Modernization of Legacy Systems: Assisting established companies in migrating or upgrading their existing infrastructures to support modern AI and Cloud capabilities.

Founded Flowr, an Integration-as-a-Service (IaaS) platform built for B2B SaaS companies requiring high-scale, reliable data connectivity. Unlike standard no-code tools, Flowr was designed with a "Type-Safe" architecture to ensure end-to-end data compatibility and high-speed execution. Key Achievements: - Type-Safe Integration Engine: Developed a proprietary orchestration engine where every data flow was strictly typed, eliminating runtime compatibility errors and ensuring robust data contracts between platforms. - High-Performance Data Exchange: Optimized data throughput by implementing high-speed exchange formats, significantly reducing latency compared to traditional JSON-based integrations. - Scalable Infrastructure: Built a customizable and monitorable platform capable of handling massive data volumes ("at scale") for enterprise-grade B2B SaaS requirements. - Full-Stack Ownership: Architected the entire system from the ground up, focusing on developer experience, system resilience, and deep observability. Technical Focus: - Data Engineering: Strongly Typed Architectures, High-speed Data Formats, Schema Validation. - Operations: Real-time Monitoring, Observability, Scalable Data Pipelines.

Overview: Key contributor to a major industrial digital transformation project, focused on migrating and optimizing predictive maintenance algorithms for the French railway fleet into the AWS Cloud. Key Responsibilities & Achievements: - Cloud Migration: Spearheaded the migration of legacy predictive maintenance applications to a modern AWS infrastructure, ensuring high availability and data integrity. - Serverless Architecture: Designed and implemented a scalable Serverless architecture using AWS Lambda, reducing infrastructure management overhead and improving execution speed. - Algorithm Optimization: Refactored and optimized Python-based maintenance algorithms within Lambda functions to improve performance and significantly reduce compute costs. - AWS Benchmarking & Strategy: Conducted deep-dive technical benchmarks of various AWS services to select the most cost-effective and performant tools for large-scale industrial data processing. - Data Engineering: Built efficient data pipelines leveraging AWS S3, Glue, and Athena for querying massive datasets, and used DynamoDB for high-speed metadata storage. Agile Collaboration: Worked within a Scrum framework, translating complex business requirements from rail maintenance experts into technical cloud specifications. Technical Stack: Cloud (AWS): Lambda, Athena, S3, Glue, DynamoDB, Cloud9, Boto3 IaC & Dev: Python, Terraform, Git, CI/CD Methodology: Scrum, Agile Analyse des besoins métier Implémentation d’une architecture Server Less

Designed, developed, and deployed a machine learning model dedicated to predicting mobile application revenue (LTV forecasting). Engineered efficient data pipelines using Python and Amazon Redshift to process high-volume user monetization data. Integrated the predictive model with Mode Analytics to visualize results and provide actionable insights for business teams.

As the sole developer on the project, I was responsible for the end-to-end design, development, and successful production launch of a social networking platform for freelancers. I transformed the initial concept into a fully functional, live product. Key Responsibilities & Achievements: - Solo Project Execution: Owned the entire development lifecycle—from the first line of code to the production deployment—managing all technical decisions and trade-offs. - Full-Stack Architecture: Architected and implemented the complete solution using the MERN stack (MongoDB, Express, React, Node.js), ensuring a scalable foundation for user networking. - Production Launch & DevOps: Successfully deployed the platform to production, configuring the server environment with Docker and establishing monitoring to ensure 24/7 availability. - Performance Optimization: Independently handled all database modeling (MongoDB) and front-end performance tuning to provide a seamless user experience. - Technical Ownership: Managed all Git versioning and CI/CD logic, ensuring a robust and bug-free release for the live user base.

As a Founding Engineer, I played a key role in building the core infrastructure of an end-to-end NLU (Natural Language Understanding) platform. I was responsible for architecting the transition from raw data to deployed production models, focusing on maximizing automation across the entire NLP lifecycle. Key Responsibilities & Achievements: - Founding Architecture: Contributed to the initial design and development of the SaaS platform, enabling users to create training sets, train models, and deploy them seamlessly. - NLP Pipeline & NER: Developed and optimized high-performance models for Intent Detection and Named Entity Recognition (NER) using Rasa and Spacy. - End-to-End Automation: Engineered automated workflows for training set creation, drastically reducing the time required to prepare models for production. - Full-Stack AI Implementation: Built the web ecosystem (React/Node/GraphQL) to visualize model performance and manage real-time NLU data. - Product Delivery: Successfully deployed a production-grade Facebook chatbot for Le Petit Paumé, serving as a real-world validation of the internal NLU framework. Technical Stack: AI/ML: Python (Numpy, Pandas, SciPy, Scikit-Learn, Rasa_NLU, Spacy) Web: Node.js, React, GraphQL, TypeScript, JavaScript

During this internship, I spearheaded the development of a custom Excel-based automation module designed to streamline the design process for summer toboggan runs. Key Achievements: Developed a selection tool in Excel that allows engineers to configure specific rail components for track layouts. Integrated Excel with AutoCAD to automate the sequential 3D modeling of track sections based on selected parameters. Optimized design workflows, resulting in a 10x reduction in design time for new track installations. Eliminated manual modeling errors and improved technical documentation accuracy. Communication inter-processus. Techos : C# / VBA / AutoCAD .NET API / Python