Tom Perrin

Design and development of complete solutions integrating data, generative AI, and full-stack web applications, from the design phase to production deployment.

▸ Design and deployment of a complete SaaS (Opifluent) independently, from architecture to production: React front-end, Node.js back-end, Supabase database (PostgreSQL), authentication, Stripe payments, Vercel deployment, monitoring, and CI/CD

▸ Integration of AI agents in production: LLMs (OpenAI, Anthropic), function calling, RAG, adaptive content generation, user behavior personalization

▸ Design of voicebots and conversational agents integrating voice recognition (Whisper, Deepgram), speech synthesis (ElevenLabs), and automatic intent detection to automate lead qualification, appointment booking, and customer support

▸ Implementation of transactional and marketing automation mailing systems (Brevo, DMARC, SPF, DKIM) with user segmentation, dynamic templates, and conversion tracking

▸ Automation of business workflows combining LLMs and no-code platforms (Make, n8n): intelligent email processing, quote generation, CRM integration, automated reporting

▸ Development of a Python quantitative framework for algorithmic decision-making: multi-strategy backtesting, automated scoring, risk indicators, 24/7 monitoring with alerts

Front-end stack: React · Tailwind CSS · TypeScript · Vercel

Back-end stack: Node.js · Supabase · PostgreSQL · Stripe · Brevo

AI stack: OpenAI · Anthropic · LangChain · RAG · Pinecone · Whisper · ElevenLabs · Function calling · Multi-agents

Data stack: Python (Pandas, NumPy) · SQL · REST/WebSocket APIs · Make · n8n

Missions and projects around data and artificial intelligence for various contexts (SMEs, entrepreneurial projects, R&D).

▸ Design of conversational AI agents and voice assistants integrating LLMs (OpenAI, Anthropic), voice recognition, and synthesis to automate business processes (request qualification, appointment booking, customer support)

▸ Development of web applications integrating generative AI to personalize user experience and generate adaptive content

▸ Implementation of large-scale data processing systems: automated Python pipelines, real-time API integration, 24/7 monitoring and alerting

▸ Advanced quantitative analyses on large volumes of data: statistical modeling, strategy backtesting, algorithmic scoring, and performance indicators

▸ Workflow automation via LLMs combined with Make / n8n: email processing, quote generation, CRM integration

Stack: Python (Pandas, NumPy) · SQL · REST/WebSocket APIs · OpenAI / Anthropic · LangChain · Make · n8n

Leading French publisher of management software (Retail, Accounting, Payroll, ERP) – Cegid supports over 750,000 companies worldwide.

Within the Retail BI team, involvement in the entire decision-making cycle for multi-brand clients (mass retail, retail specialists).

▸ Design and development of Qlik Sense dashboards for retail performance management: real-time sales monitoring, inventory management, margin and profitability analysis, commercial KPIs

▸ Architecture and maintenance of a complete Microsoft SQL Server stack: datamart modeling, multidimensional OLAP cubes (SSAS), decision support report publication (SSRS)

▸ Optimization of dimensional data models (star / snowflake) to ensure reporting performance on large volumes (several million transactions per day)

▸ Gathering business needs, designing functional specifications, and presenting to commercial management

▸ Training end-users for Qlik Sense autonomy and supporting adoption (workshops, documentation, support)

▸ Level 2/3 technical support: critical incident analysis, performance tuning, bug resolution in production environments

Stack: Qlik Sense · SQL Server · SSAS (OLAP cubes) · SSRS · SSMS · T-SQL · Dimensional modeling