Trainings
We offer trainings internally for Eki.DataPeople on various issues. Find out in this page about the trainings and tutorials we could provide access.
Our trainings can be presentations or interactive coding sessions.
Tech Session - 2025-02
Following this ambitious title,our aim is to give an overview of the evolution of Large Language Models (LLMs), from the first GPTs to today's models.
Deep Learning GenAI LLMs
Tech Session - 2025-02
Are you ready to explore a highly powerful and interpretable machine learning model? Join us for a dynamic tech session where we’ll dive into Explainable Boosting Machines (EBMs) and show you how they offer new ways of understanding and deploying machine learning in real-world scenarios.
Deep Learning Explainability
Tech Session - 2025-02
Ekimetrics x Alliance to End Plastic Waste's Prism Platform
The future of document processing is here! Join us for an exciting tech session where we'll dive into how we're leveraging advanced GenAI technologies to revolutionize document extraction and validation for the Alliance to End Plastic Waste's Prism Platform.
Deep Learning GenAI
Tech Session - 2025-01
Quantifying uncertainty in time series using conformal prediction
The Forecasting & Time Series team proposes you to attend this Tech Session on Conformal prediction in time series forecasting. During this session, we will present one of the best ways of quantifying uncertainty in time series. In other words, if you want to discover new methods (easy and more sophisticated ones) to create prediction intervals instead of predicting single values.
Deep Learning Time Series
Tech Session - 2024-12
Give sight to a blind language model
Blind large language models are dying, a new era of vision language models (VLM) has begun! In this tech session, we will share an overview of the usual tricks to give sight to a transformer-based language model.
Deep Learning GenAI LLMs
Tech Session - 2024-11
Data anonymization module of EkiParser
In this tech session, we'd like to present a project developed by the Responsible AI team : a new data anonymization module integrated into the EkiParser. The aim of this module is to identify and anonymize all personal data within documents. We'll describe the end-to-end process used to create this module, including which model was used and why, how it was evaluated, and finally its integration into the EkiParser.
Data Science NLP
Tech Session - 2024-09
We will present the latest research on LLM interpretability. It aims at “mapping the mind” of a model to better understand its behavior.
Deep Learning GenAI LLMs
Tech Session - 2024-07
Parsing solutions and EkiParser
During this tech session, we will speak about the challenges around document parsing. We will present a benchmark of the different solutions available to parse your documents. Finally we will introduce our homemade parsing solution,the EkiParser, and explain how to use it.
Data Science NLP
Tech Session - 2024-06
In this tech session, we’ll be coding a Gradio app from scratch that includes a simple chatbot, incorporating tips and tricks I've gathered from creating Gradio interfaces for LLM-based applications. This session is aimed at data scientists interested in using Gradio for POCs or prototypes of GenAI applications. We will start with a blank Python file and build a functioning Gradio GenAI application, along with a few concrete production tips.
Deep Learning GenAI
Tech Session - 2024-06
Pretraining of a Large Language Model
We will explain how Large Language Models are pretrained. We will cover the chapters 4 and 5 of the book “Build a Large Language Model”
Deep Learning GenAI LLMs
Tech Session - 2024-05
We will present the legendary paper “Attention Is All You Need” that launched the Transformer revolution and made models like GPT possible.
Deep Learning GenAI GPT
Tech Session - 2024-04
Distinguish between the direct and indirect effects of marketing levers using causality
Some marketing levers have multiple ways of affecting sales. They can have a direct effect by directly pushing the customers to purchase but also an indirect effect by activating another lever, hence affecting the customers (e.g., online media increasing the search impressions independently of the search investment). No, no, no … this is not just a matter of synergy. It’s all about marketing lever causal dependencies.
Data Science Causality
Tech Session - 2024-03
The Computer Vision Innovation Team is happy to host the next Tech.Session on gen AI for image with the tool ComfyUI! Join us for a hands-on exploration of how this technology that has been revolutionizing the generation of image.
Deep Learning GenAI Computer Vision
Tech Session - 2023-10
Discovery of LLMs and Prompt Tunning
In the rapidly evolving world of AI, you've probably heard about the incredible capabilities of Generative AI and Language Models (LLMs). After exploring the wonders of Latent Diffusion Models (LDMs), we will dive into Large Language Models (LLMs).
Deep Learning GenAI LLMs
Tech Session - 2023-10
You've probably heard of GenAI and LLMs, but what about Latent Diffusion Models (LDMs)?
Deep Learning GenAI Computer Vision
Tech Session - 2023-09
Uncertainty in time series forecasting
During this session, we will introduce the concept of uncertainty in forecasting, present popular methods and explain the different ways of evaluating them.
Deep Learning Time Series
Tech Session - 2023-03
Interpretability for Time Series Forecasting
How to assess the forecastability of a problem? How to work through a time series forecasting project? How to train your dragon model(s)? How to evaluate your model(s)? Some “tips” to run a time series forecasting model in practice.
Deep Learning Time Series
Tech Session - 2023-03
Interpretability for Time Series Forecasting
How to assess the forecastability of a problem? How to work through a time series forecasting project? How to train your dragon model(s)? How to evaluate your model(s)? Some “tips” to run a time series forecasting model in practice.
Deep Learning Time Series
Tech Session - 2023-03
We will spend some time on how to translate complex clients’ requests into a deep-learning architecture with an emphasis on making it as small as possible !
Deep Learning NLP
Reading Group Session - 2022-11
Discover FLAVA : a foundational vision and language alignment model that achieves impressive performance on all three target modalities : vision, language, and vision & language.
Deep Learning Multimodal Learning Fusion Encoders
Presentation - 2022-11
In this session, we go through state-of-the-art question answering algorithms and present a use case on how it was applied in AI for sustainability.
Deep Learning AI Question Answering
Presentation - 2022-09
In this session, we go through state-of-the-art question answering algorithms and present a use case on how it was applied in AI for sustainability.
Deep Learning AI Question Answering
Coding Session - 2022-06
Principal components Analysis in MMM
In this session, we remind of the principal components analysis, its methodology and its advantages. We then share a use case where it was applied.
PCA MMM
Coding session - 2022-03
Scraping is the fact of extracting information from a website, through a web crawler, for analysis or reuse. Learn in this coding session how to scrap vast amounts of various data using dedicated open-source tools!
Scraping BeautifulSoup Selenium
Presentation - 2022-01
Discover the critical field of interpretability for Machine Learning and how we work on state-of-the-art approaches to allow interpretability of models at Ekimetrics
Explainability Interpretability Ethical AI
Presentation - 2021-11
Detecting unknown time series events - A maritime traffic surveillance case-study
Discover an approach based on generative models for unknown event detection in time series, applied to maritime traffic surveillance using Automatic Identification System data
Time Series Maritime Traffic Surveillance Generative Models
Presentation - 2021-09
Foundation models & Prompt Engineering
GPT-3, CLIP, Codex, BERT - Foundation models are becoming the norm in many applications, from NLP to Computer Vision. What are the risks and opportunities of those models? Is it really the future of AI. Discover it yourself with your own Turing Test.
AI Research NLP
Coding session - 2021-09
Practical introduction to mathematical optimization
Turn business problems into conceptual mathematical models, find out optimal decisions using state-of-the-art solvers and apply these skills on real-world optimization problems.
Operational Research Optimization
Presentation - 2021-09
MLOps: Machine Learning in production
Machine Learning in production is a really complex discipline. Learn about the best practices and technology requirements to avoid failing in prod (input data drift, model registry, gouvernance, DevOps for ML, etc...)
Industrialization MLOps
Presentation - 2021-06
UX design guidelines for AI products
When developing products with predictive AI features, there are specific UX/UI guidelines to ensure user trust, robustness and a valuable feedback loop.
UX Design
Presentation - 2021-04
Chess has historically been one of the most important games in the development of Artificial Intelligence. Learn how to create your own chess AI using tree search algorithms, heuristics and Machine Learning.
Game AI Reinforcement Learning
Presentation - 2021-01
Cellular Automata & Artificial Life
Cellular Automata are a programming discrete model popularized by Conway and Wolfram. Discover how it can be used to model the development of Life and what concrete applications we could expect from such technology.
AI Research
Coding session - 2020-08
Object Oriented Programming for Data Scientists
Learn how Oriented Object Programming (classes) can be used in Python to accelerate your development process when you are a Data Scientist.
Python
Presentation - 2019-05
Decrypt and demystify how Netflix is using Machine Learning to power every bit of the platform, improving user experience and building a competitive advantage in the market.
AI Research