Machine Learning Solutions
Predictive analytics, classification and anomaly detection with Python, Scikit-learn, PyTorch and production-minded evaluation workflows.
Cloud data systems / applied AI
Anis Maïri
Data & AI Engineer / Microsoft Fabric / Azure
I help teams modernize data platforms, build governed analytics foundations, and turn applied AI ideas into maintainable delivery systems.
Core stack
- JavaScript
- TypeScript
- Python
- Azure Cloud Services
- 4Years of experience
- AzureCloud, AI and data platform delivery
- Contactcontact@anis-mairi.com
Expertise
Practical engineering across data platforms, AI delivery and analytics.
Clear architecture, maintainable implementation and business-readable outcomes.
Generative AI Applications
RAG, assistants and agentic prototypes using Azure OpenAI, LangChain, LangGraph, embeddings and vector search patterns.
AI & ML Deployment
Model deployment and automation with Azure Machine Learning, Azure DevOps, CI/CD pipelines and monitored delivery practices.
Microsoft Fabric & Data Engineering
Lakehouse, OneLake, ingestion, transformation and orchestration work across Microsoft Fabric, Azure data services and Databricks migrations.
Reporting & data visualization
Power BI semantic models, DAX, Power Query and dashboard delivery for business teams that need reliable, explainable metrics.
Data Strategy & Consulting
Pragmatic data strategy, governance and delivery support for teams modernizing platforms, processes and analytics operating models.
Projects
Selected data, AI and platform work
A concise view of problems solved across Fabric, Azure, Databricks, ML and Power BI.
AllMachine LearningGenerative AIFabric & Data PlatformBI & Visualization
Microsoft Fabric Migration
Migration vers Microsoft Fabric
From Azure legacy + Databricks to Microsoft Fabric
Migration d'une plateforme data existante basée sur Azure legacy et Databricks vers Microsoft Fabric, avec refonte de l'ingestion, du stockage, des transformations, de l'exposition métier et de la gouvernance data.
Impact
Modernisation d'une architecture data existante vers une plateforme unifiée, gouvernée et prête pour les usages BI et data engineering.
- Ingestion, storage and transformation redesign
- Lakehouse, semantic model and business exposure
- Governance and deployment pipeline alignment
Audit ML Forecasting
Prediction of audit closing dates
Machine learning for audit planning
Trained and deployed a regression model to predict action plan realization duration from structured audit data.
Impact
Helped audit teams anticipate delivery timelines and prioritize follow-up actions with clearer planning signals.
- Feature preparation
- Regression modelling
- Model deployment
Topic Identification
Topic identification
NLP classification for audit action plans
Classified action plans by analyzing recommendation and issue description text to identify recurring themes.
Impact
Improved portfolio-level visibility into recurring audit topics and reduced manual review effort.
- Text preprocessing
- Topic classification
- Audit portfolio analysis
Fraud and Anomaly Detection
Fraud & anomaly detection
Statistics and unsupervised ML
Built a tool leveraging statistics and machine learning to detect unusual patterns in travel and expenses data.
Impact
Surfaced suspicious behavior faster and gave reviewers a more focused anomaly investigation workflow.
- Statistical profiling
- Unsupervised learning
- Reviewer workflow support
RAG Interaction System
RAG-enabled interaction system
Azure OpenAI retrieval assistant
Developed a Retrieval-Augmented Generation tool using Azure OpenAI, ChromaDB and Streamlit, enabling auditors to interact with audit reports.
Impact
Made long audit reports easier to query, summarize and reuse through a guided conversational interface.
- Retrieval design
- Embeddings workflow
- Streamlit interface
Live Data Chatbot
Live data chatbot
Natural language over Databricks SQL
Developed a chatbot to interact with structured data in Databricks SQL Warehouse. The system dynamically queries based on user questions and retrieves information.
Impact
Reduced friction between business questions and governed warehouse data by making exploration conversational.
- Natural language interface
- Databricks SQL access
- Governed data exploration
Power BI Reporting Solutions
Power BI reporting solutions
Semantic models and dashboards
Delivered end-to-end Power BI reports for audit, merger and acquisitions, finance, accounting, green offer and sustainability teams.
Impact
Turned operational data into decision-ready dashboards with cleaner models, measures and reporting workflows.
- Semantic modelling
- DAX and Power Query
- Business reporting workflows
References
A reference network across engineering, cloud, research and consulting.
People who can speak to technical standards, collaboration and delivery context.
Contact
Useful conversations start with a concrete system, constraint or roadmap.
Available for Microsoft Fabric migration, data engineering, BI and applied AI work.