ABOUT TINYCLUES
Tinyclues’ AI-first marketing platform is built with the latest technology including ML frameworks like PySpark and TensorFlow, Kubeflow, Kubernetes, Docker or React. Our software stack is based on the Python ecosystem, natively designed for the cloud and deployed on AWS. We focus on using those technologies for what they do best.
The solution processes and analyzes hundreds of terabytes of data every day from our 100+ enterprise clients across 13 countries. It runs dozens of powerful and carefully designed Deep Learning algorithms to find the tiny clues in our clients’ first-party databases.
We believe empowering people is the most efficient way to build the best product together. The teams are organized in small amp; autonomous groups working on different business needs with agile methodologies.
Our innovative technology has led Tinyclues to be identified by leading IT analyst firm Gartner as a Vendor to Watch for digital marketing analytics and a Cool Vendor in multichannel marketing.
By joining Tinyclues, you will have an impact on one of top fast-growing start-ups with a unique AI-first vision, breakthrough predictive technology and proven global success.
(the motivation letter is mandatory)
WHAT YOU WILL BE WORKING ON
- Run state of the art predictive models in production: You will collaborate to the ever going effort needed to run in production a state of the art predictive model. The development and roll-out leads us to push the limits of KFP, PyTorch and TF.
- Improve performances: In link with Data Scientists you will work at the implementation and roll-out of various improvements: additional models, parameter tuning, new features .
- Handle 100+ configurations: Each client has its specificities in terms of data and business objectives. Therefore our predictive model is configured for each client individually. You will take part to the development of tools validating each client configuration (with relations to the code and the data) autonomously.
- Scale-up to 100+ clients: The use of resources has to be optimized so as to serve best our clients while reducing our costs. You will work on the organization of computation efforts and potentially explore new frameworks.
- Full stack teams: You will see all steps of a complete production pipeline (from vectorization to serving, passing by training) and work with a variety of profiles in our full-stack teams (Ops, Data Engineer, MLE and Data scientists)
Requirements
Must have:
- internship convention with a French Grande Ecole or a University
- the length of your internship is 5 months or more
- Fluency in Python with experience on the classic DS stack: numpy, scipy, sklearn, pandas, .
- Fluency in OOP
- basic knowledge of basic machine learning methods (linear regressions, naive bayes, PCA, kmeans)
- basic knowledge in containerization (docker, kubernetes)
- basic knowledge in algorithmic (complexity analysis, classic data structures, .)
Nice to have:
- experience with one of the Neural Network backends (TensorFlow, PyTorch)
- knowledge of Cloud environments (AWS, GCP)
- knowledge of ML frameworks (Kubeflow, Airflow)
- basic knowledge in applied mathematics (probability, gradient descent optimization.)
- relation database amp; SQL basics
- experience in data visualization
- Basic knowledge about APIs (Flask, FastAPI)
- Basic knowledge of Go
- contributing to open source software projects (Github, .)