Data & AI Engineer

Role and background

As a designer of data pipelines and real-time AI multi-agent systems, I implement advanced RAG architectures to optimize critical contact center operations and complex semantic search. With expertise in Python, LangGraph, Qdrant, and cloud orchestration on GCP via Docker, I ensure latencies under 200 ms and robust software integration. Ready to hit the ground running, I bring strong technical innovation capabilities to European companies and collaborate seamlessly in both French and English in demanding international settings.

Professional Experience

Data & AI Engineer

Since September 2025 • Business Process Outsourcing (BPO) Company – Madagascar

End-to-end design and management of a data pipeline and a real-time AI multi-agent system for a contact center with 50 agents using Vocalcom. Developed a multi-source RAG ingestion ETL pipeline using Unstructured, Gemini, and Qdrant, supplemented by a hybrid dense/sparse search engine and a Redis context cache to ensure latency below 200 ms. Orchestration of agent workflows using LangGraph, gRPC, and WebSocket for response suggestions, call summaries, and intelligent ticket escalation. Automation of call quality control via Celery, PostgreSQL, and SQL aggregation queries, significantly reducing the manual evaluation workload for supervisors. Scalable and continuous deployment on Google Cloud Platform via Docker and CI/CD, maintaining availability without service interruptions. Facilitated training workshops for client teams, promoting rapid adoption of the solution and reducing the overall request processing time by 35%.

Data & AI Developer

February 2025–July 2025 • Freelance assignment • Madagascar

Design and deployment of a comprehensive multi-source ETL pipeline for natural language querying of a legal corpus. Automated data scraping and cleaning using Python, SQL, and SeleniumBase, followed by vector indexing of over 10,000 texts via LangChain, Qdrant, and the OpenAI text-embedding-3-small model. Implementation of a RAG architecture with an embedding strategy and hybrid search to optimize the relevance of results. Development of an API using FastAPI that integrates scheduled processing via cron jobs, containerized with Docker, and deployed on OVH via a CI/CD pipeline. Robustness and maintainability ensured through rigorous testing, structured logs, and detailed technical documentation. This project improved query accuracy by 40% compared to traditional lexical search, while guaranteeing a median latency of less than 300 ms.

Full-Stack Developer

June 2024–December 2024 • UEAS — Academic Achievement Assessment Unit • Madagascar

Design and deployment of a web application for centralizing academic data using React.js, Node.js, Express, MongoDB, and PM2, incorporating advanced aggregation queries and quality control mechanisms. Simultaneously implemented a comprehensive Security Operations Center (SOC) using Zeek, Suricata, and Splunk for network traffic capture, proactive log analysis, and intrusion detection. Led change management efforts with end users to ensure a smooth transition to the platform. These achievements have made it possible to secure the network infrastructure in the long term, ensure full adoption of the tool by the teams, and reduce data collection times by nearly 60%.

Full-Stack Developer

August 2023–December 2024 • MyTsenaTech • Madagascar

End-to-end design and deployment of a bus ticket booking platform, including user interfaces and the back-office management system. Developed the full-stack architecture using Next.js, TypeScript, Express, and PostgreSQL, leveraging Prisma for relational modeling and Zod for strict data validation. Full integration of the MVola mobile payment API via OAuth, covering transaction initiation, tracking, and securing financial flows. Implementation of advanced access management using JWT tokens, role-based control, and optimization of asynchronous requests with React Query. Delivery of a fully operational payment solution and maintenance of complete data integrity in production, with no reported incidents.

Skills

Hard Skills: RAG Architecture & Multi-Agent Systems (LangGraph, LangChain), ETL Pipelines & Real-Time Data Ingestion, Vector Databases & Hybrid Search (Qdrant, Redis), Backend & API Development (FastAPI, Node.js, Python), Cloud, Containerization & CI/CD (GCP, Docker)

Soft Skills: Change Management & User Training, Performance Optimization & Latency Optimization, Proactive Problem-Solving for Complex Issues, Commitment to Security & Data Reliability, End-to-End Project Management and Delivery

Strengths: French (fluent), English (business level), Expertise in RAG, vector databases, and multi-agent systems, Dual expertise in Data/AI and Full-Stack development, Proficiency in real-time ETL pipelines and cloud architectures

Degrees and Education

Master’s in Computer Science

2022–2024 • École Nationale d'Informatique (ENI), University of Fianarantsoa

Specialization in General Computer Science

Bachelor’s Degree in Computer Science

2018–2022 • National School of Computer Science (ENI), University of Fianarantsoa

Specialization in General Computer Science

At a glance

Field of expertise

IT & AI

Experience

3 years of experience

Working languages

French · C1English · B2

Tools and software in use

JavaScript — 4 yearsPython — 3 yearsSQL — 3 yearsTypeScript — 3 yearsPostgreSQL — 3 yearsMySQL — 3 yearsNode.js — 3 yearsExpress — 3 yearsMongoDB — 3 yearsWebSocket — 3 yearsPrisma — 3 yearsZod — 3 yearsQdrant — 3 yearsLangChain — 3 yearsOpenAI API — 3 yearsHugging Face — 3 yearsGit — 3 yearsReact.js — 3 yearsSeleniumBase — 3 yearsPHP — 2 yearsApache Airflow — 2 yearsApache Kafka — 2 yearsRedis — 2 yearsFastAPI — 2 yearsNestJS — 2 yearsLaravel — 2 yearsGraphQL — 2 yearsgRPC — 2 yearsSQLAlchemy — 2 yearsPyTorch — 2 yearsScikit-learn — 2 yearsPinecone — 2 yearsLangGraph — 2 yearsVertex AI — 2 yearsMLflow — 2 yearsn8n — 2 yearsPM2 — 2 years

Markets covered

SwitzerlandFranceBelgiumCanada