Stéphane KPOVIESSI.
AI Systems & Data Infrastructure Engineer
Designing scalable multi-agent platforms, deterministic big data pipelines, and high-performance execution runtimes. Dedicated to transforming volatile models into production-grade pipelines.
Engineering Profile & Interests.
AI Systems & Data Infrastructure Engineer specializing in architecting distributed Big Data pipelines and orchestrating autonomous multi-agent systems. My expertise focuses on the data lifecycle core, ensuring data integrity, transformation, and leverage through robust and scalable infrastructures.
My engineering practice is rooted in designing complex data environments. I have driven the modernization of critical pipelines at Allianz France (SAS to PySpark/Databricks migration using Medallion architecture) and developed generative vision systems (GANs) at Groupe Sylvagreg. Deeply interested in agentic AI, I build advanced orchestration systems using LangGraph and the Model Context Protocol (MCP) to bridge the gap between research and operational systems.
Systems Architecture Rule
FROM METRIC TO MACHINE"I don't just train models. I architect the infrastructures that allow data and AI to operate reliably, ensuring stream quality and system consistency at scale."
Data Platform Engineering
Architecting highly reliable Bronze/Silver/Gold Medallion pipelines, distributed PySpark systems, data quality validations, and transactional storage structures for massive corporate datasets.
Autonomous Agentic Systems
Designing multi-agent state machines, context-aware RAG pipelines, LangGraph and Pydantic AI orchestration, Model Context Protocol (MCP) servers, and isolated runtime sandboxes.
Computer Vision & Generative AI
Developing real-time video processing pipelines, YOLO object tracking, image segmentation with SAM, GAN optimization with PyTorch, and procedural 3D environment rendering.
Automation & Data Performance
Developing automated ingestion tools, optimizing processing workflows, and building analytical data structures to transform raw streams into actionable business insights.
Operational Releases & Experience.
Data Engineer
Led the end-to-end modernization of critical commercial property and liability data pipelines. Successfully migrated legacy SAS codebases to a distributed PySpark architecture on Databricks with absolute parity.
[MODERN MEDALLION ARCHITECTURE]
- ↳Designed and built a modular Bronze/Silver/Gold pipeline structuring portfolios, asset capitals, and operational emissions.
- ↳Restructured transactional staging to store data efficiently in optimized, compressed Parquet tables, improving downstream read times.
[LEGACY DECONSTRUCTION]
- ↳Deconstructed, reverse-engineered, and documented over 15,000 lines of legacy SAS scripts across 19 interconnected business modules.
- ↳Preserved 100% of complex insurance accounting and risk business logic during translation to PySpark (Python 3.9+).
[FUNCTIONAL PARITY & QUALITY ASSURANCE]
- ↳Designed an automated schema validation framework that rejects corrupted or structural-shifting data points at ingestion.
- ↳Implemented functional parity testing pipelines to run SAS and PySpark nodes side-by-side, validating and correcting data down to the float decimal.
[GEOSPATIAL DATA CLEANSING]
- ↳Designed an ETL pipeline for raw coordinates: parsing and reconstructing GPS locations, validating postal address syntax, and resolving outliers.
- ↳Reduced address parsing false positives, establishing a deterministic baseline for geographic risk reporting.
Co-Founder & CTO
Co-founded and engineered an AI-orchestrated collaborative workspace platform. Led the technical team, designed backend infrastructure, and prototyped agentic engines.
Machine Learning Engineer
Researched and built a generative computer vision solution to construct printable, geometrically coherent 3D models directly from single 2D structural images.
Business Data Analyst
@LYNE ET FRERES SARL
Analyzed commercial sales, import/export records, and maritime shipping logistics to optimize supply chain inventory flow and commercial performance.
Community & Leadership Changelog
Vice President
@JUNIA Alumni
- ↳Coordinated committee operations, organized networking summits, and supervised creative design and event production teams.
- ↳Managed active student and alumni community integrations in partnership with the student union.
Head of Communications
@JUNIA Alumni
- ↳Orchestrated professional alumni networking events, managed digital community engagement, and synchronized student-alumni experience sharing.
Technical Case Studies & Systems.
- Established an extensible, secure agent execution runtime.
- Engineered pluggable skills that allow agents to read local files and query external APIs under strict user permission grids.
- Reduced system-to-model overhead by employing lightweight gRPC channels.
- Automated complete daily technology summarization and research discovery pipelines.
- Engineered robust multi-agent coordination with designated debate and editing loops, producing high-fidelity technical reports.
- Aggregated and cleaned multi-channel commercial records into a unified analytical data mart.
- Designed cohort segmentation trackers and sales trend telemetry dashboards, reducing reporting latency.
- Optimized retail stock calculations, providing real-time data support for downstream inventory planning.
- Prototyped procedural vegetation structures displaying low vertex overhead.
- Created a robust coordinate rendering pipeline for simulated training environments.
- Automated soccer game recording parsing, generating actionable movement charts.
- Provided amateur clubs with high-fidelity coordinate metrics previously limited to professional enterprises.
- Optimized solar grid storage dispatch simulations.
- Designed a robust feature parsing loop that handles missing atmospheric readings safely.
AUXILIARY EXPERIMENTAL NODES
Created a high-speed CLI/TUI developer tool that maps complete local folders, generates accurate tree models, and structures complete project files into an optimized markdown digest.
Developed a clean, service-oriented backend API in Go, showcasing modular routing patterns, secure session handling, and transactional database mapping.
Designed and coded a high-density, multi-panel desktop control interface for the Sylva3D rendering engine, displaying parameter dials, scene trees, and engine status indicators.
EXPLORE COMPLETE CLUSTER NODES
I actively contribute to open-source software, testing local sandboxes and structural parsing strategies. Explore my global GitHub organization for full implementation scripts.
Technical Specifications Grid.
Data Engineering
READY FOR EXECUTIONAI Systems & Orchestration
READY FOR EXECUTIONComputation & Core
READY FOR EXECUTIONAcademic Foundations.
Master of Science in Engineering (Diplôme d'Ingénieur)
2023 - 2026MASTER'S DEGREE (BAC+5)JUNIA ISEN Lille // Big Data & Data Science
Elite French Graduate School of Engineering. Advanced curriculum specializing in massive data engineering, statistical modeling, distributed systems, and production machine learning.
Bachelor of Science in Digital Engineering
2020 - 2023BACHELOR'S DEGREE (BAC+3)JUNIA ISEN Rabat // Electronics, Computer Systems & Digital Sciences
Intensive undergraduate engineering program covering foundations of computer science, advanced algebra, embedded computing, and digital systems.
Classe Préparatoire MPSI / MP
2020 - 2022Marie Stella // Advanced Mathematics, Physics & Computer Science
Highly selective, intensive two-year undergraduate program preparing candidates for the French Grandes Écoles entrance examinations.
Verified Registry.
Microsoft Certified: Azure Data Scientist Associate
ISSUER: Microsoft (Coursera Verify LU2ZQ972BWNF)
Data Analysis with Python
ISSUER: FreeCodeCamp (Verify Certification: Amiche)
Supervised Machine Learning
ISSUER: DeepLearning.AI (Verify MWJS4Q3S2C3L)
Artificial Intelligence A-Z
ISSUER: Udemy (Verify UC-f0f26f78-fc5e-4cd5-a5f8-7c7f64680774)
"Continuous alignment with changing cloud paradigms, mathematical operations research, and automated workflow state machines."
Initiate Connection Gateway.
Open to collaborative development on autonomous agentic systems, PySpark migration audits, microservice designs, or other challenging engineering initiatives.