Career
Summary
Data Engineer with 2+ years building and operating production data platforms on Microsoft Azure. At JACQUEMUS I own over 20 enterprise pipelines, built on a medallion lakehouse architecture in Microsoft Fabric and orchestrated with Fabric Data Factory, Spark notebooks and live Quarkus web services. I lead FinOps and Azure administration (PowerShell automation, IAM, compute rightsizing, shared Spark pool configuration) and standardized Datadog observability across the platform and its data flows.
Experience
-
Data Engineer · JACQUEMUS
Paris, France · Aug 2024 - Present (2 years 3 months)
JACQUEMUS is a French luxury fashion house selling ready-to-wear, leather goods, and accessories through its own e-commerce, its boutiques, and wholesale partners.
Part of the Data Platform team, owning the retail order path end to end across OMS (order management system) and OCI (omnichannel inventory) for e-commerce, boutique, and wholesale channels. I design and operate the surrounding financial, product, pricing, warehousing, and invoicing pipelines on a Microsoft Fabric medallion lakehouse, while leading platform FinOps and cloud cost governance.
Data Engineering
- Built and maintained over 20 enterprise pipelines across Salesforce, ORLI/WMS, REST APIs, SFTP file shares, Azure Data Lake, and Azure SQL (SQL Server). Impact: The analytics, logistics, e-commerce reporting, and inventory-management teams work from these flows, so a stalled pipeline delays morning commercial reporting and inventory synchronization.
- Engineered the platform on the modern Azure data stack: a medallion lakehouse on Microsoft Fabric, orchestrated with Fabric Data Factory and transformed with PySpark notebooks, landing order, financial, product, and pricing data across bronze, silver, and gold Delta layers. Impact: Establishes an enterprise single source of truth with structured data layers, accelerating downstream Power BI analytics and cross-departmental reporting.
- Integrated live Quarkus REST API services behind Azure Application Gateway alongside scheduled batch jobs for real-time order and inventory events. Impact: Bypasses batch refresh windows for high-priority telemetry, moving time-sensitive retail updates directly over low-latency API endpoints.
Cloud & FinOps
- Reduced Azure infrastructure spend by €1,400 per month by automating development-environment shutdowns with PowerShell runbooks in Azure Automation Accounts, managing Azure IAM access controls, matching compute size to each environment, and redesigning shared Spark pools. Impact: Delivers a recurring monthly saving on the platform budget with zero SLA impact on morning business reporting pipelines.
Monitoring & Observability
- Instrumented ingestion workloads with Datadog APM and custom monitoring dashboards across Azure Data Factory, Microsoft Fabric, and Spark execution pools. Impact: Provides end-to-end alert coverage across 150+ production jobs, surfacing silent job or API failures before business-hours reporting windows.
- Authored platform architecture runbooks, disaster recovery procedures, and operational playbooks for the 150+ production pipelines and the retail order path. Impact: Empowers on-call engineers to recover failed runs directly from standard playbooks, shortening mean time to recovery (MTTR) during incident responses.
-
Data Engineer · OLIVESOFT
Sfax, Tunisia · Feb 2024 - Jul 2024 (6 months)
OLIVESOFT is a Paris-headquartered data consultancy delivering data engineering, integration, and BI services to over 20 clients, primarily in luxury retail, spanning platform delivery, consulting, and end-to-end operational support.
Engineered the production customer-feedback integration pipeline for the Kenzo Paris / LVMH account, building a 3-layer event-driven MuleSoft architecture on CloudHub between Salesforce Service Cloud and Diduenjoy.
- Architected an asynchronous 3-layer MuleSoft API-led connectivity framework (System and Process APIs) on MuleSoft CloudHub, using DataWeave 2.0 to transform Salesforce order event triggers into Diduenjoy survey payloads. Impact: Decoupled agent workflows and achieved a 42% customer survey response rate in the initial monthly report, establishing an NPS 48 baseline toward Kenzo's 55 NPS target.
- Implemented RAML 1.0 REST contracts, Anypoint MQ queues, and a 5-stage exponential backoff retry policy distinguishing transient retries from permanent routing to Dead-Letter Queues (DLQ). Impact: Guaranteed zero feedback data loss during third-party API outages, with automated CloudHub PaaS notifications handling failure alerts.
-
Machine Learning Engineer · REGIM Lab
Sfax, Tunisia · Jun 2023 - Aug 2023 (3 months)
REGIM-Lab (REsearch Groups in Intelligent Machines) is the applied-research laboratory at the National Engineering School of Sfax (ENIS), developing intelligent systems, machine learning models, and sensor data architectures.
Engineered a research prototype for secure, decentralized energy monitoring, ingesting smart-meter time-series telemetry from Australia's Ausgrid dataset to benchmark regression anomaly detection models and log tamper-proof cryptographic audit trails.
- Ingested and preprocessed 1 year of 30-minute interval telemetry (~5.2M records) across 300 solar households with Pandas and NumPy, delivering live metrics via a Flask, Jinja2, and JavaScript dashboard.
- Engineered time-series features and benchmarked Scikit-learn regression models to flag meter anomalies, writing cryptographic hashes to Solidity smart contracts via Web3.py on a local Ganache testnet. Impact: Formed the core data architecture and empirical benchmarks for a peer-reviewed research paper published in Computers & Industrial Engineering (Elsevier, 2025).
-
OEM ENGINEERING S.A.R.L
Sfax, Tunisia · Jun 2022 - May 2023 (1 year)
OEM ENGINEERING is a Canadian industrial engineering company headquartered in Sherbrooke, Quebec, with its engineering site in Sfax, Tunisia, building high-speed material-sorting systems that identify metal alloys inline from visible near-infrared (VNIR) camera signals.
Engineered the real-time signal processing and inline material classification stack across two roles, building high-speed C++ filtering at ~2,000 FPS and offline XGBoost spectral model benchmarks.
Machine Learning & Statistical Engineer
Sep 2022 - May 2023 (9 months)
Real-Time Classification
- Integrated the deployed C++ Savitzky-Golay filtering path into inline metal-alloy classification, sustaining approximately 2,000 frames per second (<0.5 ms per frame latency budget) for real-time sorting decisions. Impact: Alloy identification runs at full conveyor speed, meeting the strict hardware throughput budget required for inline sorting.
Model R&D
- Engineered spectral feature baselines and evaluated offline XGBoost models using Scikit-learn and NumPy across a ~10,000+ VNIR sample dataset, achieving 94.7% accuracy across multi-class metal alloys while maintaining production separation due to offline latency constraints.
Signal Processing Engineer
Jun 2022 - Aug 2022 (3 months)
- Delivered a 100× signal-preprocessing speedup by replacing a C# Savitzky-Golay implementation with optimized C++ filtering using fixed in-memory parameters and GCC optimization, while retaining the same filter behavior for inline sorting. Impact: Preprocessing stopped being the bottleneck, and this is the path that went inline the following autumn, where the filter has to keep pace with the camera rather than catch up behind it.
Education
-
Computer Science Engineer’s Degree · National Engineering School of Sfax
Sfax, Tunisia · 2021-2024 (3 years)
Computer science engineering program specializing in Data Engineering, Distributed Systems, Cloud Architectures, and Applied AI.
Graduation Thesis
- Subject: fault-tolerant integration between a Salesforce service desk and an external customer-feedback platform, studied over a six-month final-year placement at OLIVESOFT.
- The argument: queued, asynchronous exchange keeps the service desk working when the downstream platform is slow or unavailable. The report set out the architecture, the failure cases it covers, and the trade-offs it accepts.
- Outcome: a written thesis and a running production integration, defended for the degree in 2024.
Data & Systems Specialization
- Data engineering & cloud: big data processing, database management systems (SGBD), business intelligence, parallel & distributed computing, and cloud architectures.
- Applied AI: machine learning and computer vision.
- Software craft & systems: DevOps, microservices, industrial blockchain, and clean coding.
Student Clubs & Training
- Training Lead, Securinets ENIS: ran the club's technical training track, 6 sessions a year (3 per semester), from x86 assembly through hands-on reverse-engineering labs to web exploitation and SQL injection, open to club members and to any ENIS student who turned up.
- Finding the people to teach it: brought in senior students and working engineers for the topics the club could not cover itself, and built the term's schedule around who was available.
- Member, IEEE Student Branch ENIS: part of the group running the introductory programming series (CS fundamentals, Python, C++ for competitive programming) .
-
Bachelor of Engineering (Mathematics and Physics) · Higher Institute of Applied Sciences and Technology of Mahdia
Mahdia, Tunisia · 2019-2021 (2 years)
The Tunisian national engineering preparatory program, MP track: intensive mathematics and physics, with admission to the National Engineering School of Sfax (ENIS) decided by competitive national ranking.
- Advanced mathematics: real and complex analysis, linear algebra, matrix theory, differential equations, numerical methods, and probability theory.
- Physics & computational foundations: classical mechanics, electromagnetism, thermodynamics, wave physics, and algorithmic problem solving.
Certifications
Vendor-certified across the data engineering stack: cloud data platforms, data orchestration, data integration, and observability.
-
Microsoft
-
Astronomer
-
MuleSoft
-
Talend
-
Datadog
Online Courses
Continuous learning to master computing fundamentals and emerging technologies, understanding how new tools interface into modern data architectures.
-
Anthropic
-
OpenCV
-
Coursera
-
Udemy
-
DataCamp
-
Scrumstudy
Volunteering
Crisis response, before any of the rest of this.
-
Jeunes Ingénieurs de Djerba (JID)
Djerba, Tunisia · Orientini · Jul 2023 (1 month)
- Guided secondary school students through engineering fields and the careers each one leads to.
-
Jeunes Ingénieurs de Djerba (JID)
Djerba, Tunisia · Orientini · Jul 2022 (1 month)
- Guided secondary school students through engineering fields and the careers each one leads to.
-
Tunisian Red Crescent · Houmt Souk branch, Djerba
Djerba, Tunisia · COVID-19 response · Mar 2020 - Jun 2021 (1 year 4 months)
- Helped organise aid distribution and assisted with crowd coordination.
- Collected and distributed essential supplies, and supported elderly residents.