Isaac Yimgaing Kuissu

At Thales, a global leader in aerospace, defense, and security technologies, I assumed the role of an LLM & MLOps Engineer. My main responsibilities included: ◽ Design and Deployment of Fine-Tuning Pipelines for LLM Models: Developed a robust and scalable production pipeline using MLflow. This infrastructure significantly improved the efficiency and accuracy of LLM models, in line with Thales' data security and confidentiality requirements. ◽ Development of Continuous Evaluation Pipelines for LLM Models: Established an innovative architecture for continuous evaluation and retraining of models, ensuring their reliability and optimal performance in various production scenarios. ◽ Architecture for the Deployment of Portable Solutions: Designed and implemented a deployment architecture that respects strict data confidentiality and security constraints while ensuring maximum portability and adaptability of solutions. ◽ Enhancement and Development of an Advanced Chatbot: Utilized a RAG architecture combined with VectorDB to build and enhance a high-performing chatbot, contributing to the optimization of user interactions and the efficiency of internal processes. In addition to these technical tasks, I also collaborated closely with international and multidisciplinary teams to integrate AI solutions into a broader range of applications and actively participated in the development of MLOps strategies for better evaluation, efficient deployment, and continuous monitoring of models in cloud environments. technologies: Azure ML, Databrick, Vector DB, Huggingface, LLM models

◽ Analyze the business need, design, deploy and continuously improve contextual targeting models to automatically categorize +2M media content (text, image, video) per day, in order to target and automate digital advertising while guaranteeing ROI to brands and a good user experience on websites. ◽ Manage the data science division, manage the project schedule in accordance with the business roadmap, and popularize AI within the teams. ◽ Building and implementing a media content recommendation system based on the similarity of media content in the database via GPT-3 ◽ Construction and production of a tool for optimizing the ROI of advertising campaigns based on ChatGPT ◽ Construction of a multi-channel competitive intelligence monitoring tool based on ChatGPT

Factonics is a French startup, publisher of a secure platform of ready-to-use AI services for business departments. As a Data Scientist, my role in the team is to: ◽ Contribute to the evolution of Factonics' core platform by building new use cases ◽ Contribute to the qualification of needs and feasibility study for new accounts ◽ To the construction and integration of new use cases on the platform via Machine Learning (Deep Learning, NLP, NLU, Image Processing, Segmentation,...) ◽ To the production launch of AI solutions developed in an AWS Cloud environment ◽ Contribute to the implementation of DevOps & DataOps practices (agile SCRUM, Jira, Git, CI/CD,..) ◽ Manage a small team of data Scientists working on platform evolutions ◽ Support the growth of the team by interviewing data science candidates ◽ Mentoring data scientist interns ◽ Leading workshops and workshops to popularise AI

Mentoring is above all about sharing knowledge, passing on one's passion, but also helping students to retrain for a job they love. This is the role of pedagogical and technical coach that I play at OC. My mentoring courses: ◽ Data & Machine Learning ◽ AI Engineer

In a world of Data that is constantly evolving, I accompany students in the discovery of the Data professions, their interactions, opportunities, thus helping them to better perceive the Data ecosystem and its challenges.

Within the IALab team, our goal was to respond, via AI and Machine Learning, to the needs and problems of all Bpifrance entities. During this experience, I was able to: ◽ Lead 5+ workshops and workshops to popularize AI to the business lines ◽ Take part in the deployment of an AWS Cloud infrastructure Examples of applications developed: ◽ Design, development of an AI to support 850K+ entrepreneurs in their life moments through AI (project selected and presented at AI for finance). Construction (MVP) of a product and service recommendation engine based on customer knowledge (CRM, website, events,..) allowing to model a 360 vision of a customer/prospect and thus to propose the right services at the most appropriate time (Scikit Learn, H20, Python, Git, SQL, SCRUM) ◽ Design, development of an AI for automatic detection of entities (company name, fund raising, theme,..) and sirenisation of national and international press articles (300+ per day) in order to facilitate the detection of new opportunities by the business managers (Elastic search, Kibana, Docker, NLP) ◽ Analysis and categorisation of business risks based on SEC (Securities and Exchange Commission) reports collected via scraping (350GB+ of data), then preprocessed and modelled (NLP, NLU, LDA). Set up a continuous collection pipeline with Airflow and visualisation (EC2, Redshift, S3, scrapy, plotly dash) ◽ Construction of an ETL for extracting and cleaning data from Oracle Eloqua Marketing to model user paths on Bpifrance websites (Market Basket Analysis, Hierarchical Clustering, DBSCAN, HDBSCAN) ◽ Construction and implementation of a multi-source & multi-channel log categorisation solution, automatic detection of errors and anomalies through log analysis via Machine Learning (Xgboost, Random Forest, Flask)

Within the Dream team, we were tasked with implementing a method of predicting electricity consumption on an hourly basis in order to help the purchasing teams refine their daily purchase orders. As a Data Scientist, my role in the team was to: ◽ Audit the Machine Learning algorithm implemented via technical and business metrics. ◽ Propose and implement different other modelling to meet the use case (Deep Learning, time series,..) ◽ Carry out a benchmark of the models obtained (performance, advantages & disadvantages) Tools: Tensorflow, Scikit Learn, Python

Provide our insurance clients with analysis, reports and recommendations for strategic planning across all business units. As a Data Analyst, my role in the team was to: - Analyze/Audit data to transform it into actionable information via the Power BI ETL - Propose and implement the company's Data-Driven strategy - Propose business dashboards to facilitate decision making Tools: Python, Power BI, Power point

Co-pedagogical and technical trainer of students on the bilingual Business Intelligence and Big Data course. Tools: Power BI, Excel, Hadoop, Python