Wissam El Jablaoui

- Technologies used: Python, AppScripts, LangSmith - Developed automation scripts and API integrations, including documentation, to reduce manual data entry and streamline ingestion workflows, resulting in up to a 60% reduction in data entry time. - Collaborated with prompt‑engineering and product teams to refine prompt templates and guardrail definitions, establishing a standardized prompt library and response‑validation procedures for model interactions. - Enhanced an existing LLM-based agent to support file ingestion and structured output: built a multi-agent system that indexes large document repositories by generating searchable summaries, then answers questions by searching source files and extracting quotes with page numbers.

- Technologies used: Python, Flask, Gemini API, Firebase (Auth, DB & Storage), Next.js (React), Rest API, Git, Docker, GCP . - Developed a bilingual (Arabic/English) exam-generation and document-handling web application using React for front-end, Firebase for storage and DBs, Flask for the backend with Google’s Gemini API, enabling dynamic data-driven content creation. - Redesigned API workflows to reduce LLM hallucinations by segmenting requests by question type and difficulty, implementing stop mechanisms, and optimizing response validation. - Improved system reliability through asynchronous processing, rate limiting, and retry logic with exponential backoff and model fallback. - Containerized services with Docker and deployed on Google Cloud Run to ensure scalable, efficient, and reproducible environments for rapid testing and deployment.

- Technologies used: Power BI, SQL, Python, Git. - Mentored and coached a cohort of 13 students through the IA Data Analyst curriculum, focusing on developing analytical skills and leveraging industry-standard tools. - Designed and delivered engaging lessons, practical exercises to prepare students for Microsoft PL-300 certificate. - Upgraded educational materials for Machine Learning, with a focus on scikit-learn and Deep Learning using PyTorch, to provide students with hands-on experience. - Designed an AI-powered quiz generator using Langchain and Hugginface’s open-source models to streamline the creation of assessments, enabling coaches to deliver customized quizzes faster and more effectively. - Taught a basic Data Engineering curriculum covering tools like Docker, Apache Spark, Kafka, BigQuery, dbt, and Mage for workflow orchestration. The program included hands-on modules in data ingestion, warehousing, analytics engineering, and real-time streaming. - Assisted a Master's student in Cryptology & Cybersecurity with designing and implementing a machine learning model to predict malware presence in files and URLs. Supported data preprocessing, feature engineering, model selection, and evaluation.

- Technologies used: Python, Git, Docker, SQL, Hugging Face, MLflow, FastAPI, Langchain, Fivetran, Looker Studio, Power BI, Zapier, dbt, Census. - Assisted students with their learning challenges including practical challenges and theoretical knowledge. - Refined and updated curriculum content to enhance delivery and facilitate better student understanding.

- Technologies used: Jupyter Notebook, Python, R, Excel VBA. - Implemented a landscaping tool using Excel VBA and dimension reduction techniques to guide the product management team on how to improve the product based on sensory and physical attributes collected among Ab-inBev ’s and competitor’s products. - Applied web scraping and NLP techniques to extract valuable insights from textual data in surveys, leading to actionable insights for the management. - Developed a Python script and implemented survival analysis techniques to gain insights on the impact of physical/chemical attributes on the product’s shelf-life index and providing recommendations to the team.

- Technologies used: Power BI, MS Project, Visio, Google OR tools. - Developed project plans for two retail showrooms to keep track and report the project progress and costs directly to the General Manager on a weekly basis. - Developed Standard Operating Procedures and Visio Charts for Omnichannel processes, serving as official documents for training staff members. -Implemented Bitrix24 CRM to automate the sales cycle starting from lead collection (through marketing campaigns) to conversion of the latter into deals, boosting sales by 18%. - Led and trained 43 employees from different departments in order to properly collaborate on Bitrix24. - Utilized Python & Google OR tools to work on optimizing resource allocation and delivery pathing for B2B deliveries (FMCG products) using Linear Programming (Operations Research – Capacitated Vehicle Routing Problem).