Malek Sahlia

Agentic RAG Chatbot For Banks: Developed an Agentic RAG chatbot solution for the banking sector using Azure AI services, enabling intelligent document-based conversational assistance. Built an end-to-end document ingestion pipeline leveraging Azure Blob Storage, Event Grid, Azure Document Intelligence, Azure AI Search, and OpenAI embeddings (text-embedding-ada-002). Implemented conversational memory and caching mechanisms using Redis to optimize chatbot performance and response latency.

AI Engineer (FinTech Industry) - Value Digital Services - Tunisia, Hybrid

(2026-01)

• Agentic RAG Chatbot For Banks: Developed an Agentic RAG chatbot solution for the banking sector using Azure AI services, enabling intelligent document-based conversational assistance.
• Built an end-to-end document ingestion pipeline leveraging Azure Blob Storage, Event Grid, Azure Document Intelligence, Azure AI Search, and OpenAI embeddings (text-embedding-ada-002).
• Implemented conversational memory and caching mechanisms using Redis to optimize chatbot performance and response latency.
• Integrated MLflow for prompt/version tracking and used RAGAS for evaluating retrieval and generation quality.
• Monitored and improved AI workflows using Laminar observability tools for tracing and performance analysis.
• AI-Driven Client Portfolio Analysis: Developed an AI-powered portfolio analysis system for banking clients using tabular financial data and fixed banking KPIs.
• Leveraged LLMs to generate automated portfolio insights, risk interpretation, and personalized financial analysis for client portfolios.
• AI Data Engineering & Automation Solutions: Developed web scraping pipelines to collect and structure large-scale data from online sources.
• Deployed AI services and APIs using FastAPI, Docker, and asynchronous task processing with Celery.
• Designed and orchestrated data pipelines using Airflow DAGs for automated data ingestion and processing.
• Performed large-scale analysis of social media reviews and user feedback using LLM-based NLP techniques.
• Built data processing workflows and integrated AI models into production environments.

AI Team Lead (HRTech Industry) - Fabskill - Tunisia, Hybrid

(2025-06 - 2026-01)

• Lead AI team developing CV parsing and matching solutions using RAG, LLMs, OCR, and Milvus.
• Deploy APIs using FastAPI and Docker for scalable AI services.
• Manage Git workflows and collaborate cross-functionally to deliver AI-powered recommendation features.
• Drive improvements in parsing accuracy and system performance.

Data and Machine Learning Engineer (HealthTech Industry) - Knowlepsy - Tunisia, Remote

(2023-12 - 2024-11)

• Contributed to building models for epileptic seizures triggers.
• Built Dash + PostgreSQL dashboards, reducing manual reporting.
• Designed scalable ETL pipelines on Azure (10K records/sec, 96% availability).
• Documented models and added unit tests for reliability.
• Worked on time series (wearables), LLMs (chat/forms), and Computer Vision (meal/activity detection).

Data Science Engineer | AI-Driven Hospital Resource Optimization - Institute of Intelligence and Data (IID), Laval University - Canada, Hybrid

(2023-03 - 2023-08)

• Developed predictive AI models to estimate patient diagnosis severity and hospital length of stay, ensuring fairness across gender and other sociodemographic factors.
• Optimized hospital admission workflows, achieving a 12% improvement in resource utilization and patient throughput at IUCPQ hospital.
• Conducted systematic evaluation of AI model performance and bias to ensure actionable, reliable results for clinical and operations teams.
• Delivered technical reports and presentations to stakeholders to support deployment of AI-driven solutions in production environments.

Data Scientist | Extraction and Conversion of Invoices to CSV - AxeFinance - Tunisia, On-Site

(2022-07 - 2022-08)

• Automated data extraction from financial statements using Docker and OCR, reducing manual effort by 66%.
• Implemented a scalable pipeline for table detection, text parsing, and CSV conversion.
• Orchestrated services with Docker Compose, enabling scalable processing of financial data.

PhD - Industrial Engineering - Polytechnique Montréal (2026-08 - 2029-08)

Master's in Research - Intelligent Systems - National School of Computer Science (ENSI) (2021-09 - 2023-09)

National Engineering Degree - Computer Science - National School of Computer Science (ENSI) (2020-09 - 2023-09)

Undergraduate Degree - Mathematics and Physics - Preparatory Institute for Engineering Studies (2018-09 - 2020-08)