ML Engineer Intern

Since 2014, Wiremind has positioned itself as a tech company transforming the world of transport and events with a 360° approach combining UX, software, and AI.

Our expertise lies primarily in optimizing and marketing our clients' capacity. We work on various projects such as ticket forecasting and pricing, 3D optimization of air freight or scraping competitor prices. Our applications are the preferred tool of companies such as SNCF, United Airlines, Qatar Airways or even PSG to visualize, analyze and optimize their capacity.

Dynamic and ambitious, we strive to maintain our technical DNA which is the engine of our success. The company, profitable and self-financed since its creation 10 years ago, is mainly composed of engineers and experts and currently supports the growth of our business model based on "software-as-a-service" solutions.

At Wiremind, the Data team is responsible for developing, monitoring, and evolving all ML-powered forecasting and optimization algorithms used in our Revenue Management systems. The team consists of Machine Learning Engineers, who develop and monitor ML models, and Data Engineers, who ensure smooth integration of these models into our software.

Our algorithms are divided in 2 parts:

• ML models to forecast unconstrained demand, capacity and overbooking (e.g. boosted trees, deep learning), trained on historical data in the form of time-series
• Constrained optimizations problems solved using linear programming techniques

CARGOSTACK is an end-to-end SaaS solution for airlines to manage their cargo business. It combines a Cargo Management System (CMS) with optimization modules leveraging AI models to provide forecast to support users when taking decisions such as setting the right level of capacity for sale (in kg and m3), the right price, the right overbooking level, or the best pallet build-up (think of a 3D-Tetris optimizer)

As a Machine Learning Intern, with support from more senior ML Engineers, you will leverage state-of-the-art AI/ML methods and ironclad validation processes to deliver robust, interpretable prediction systems. Your responsibilities will include:

• Designing/training of new models (decision trees or neural network) to improve our current models in production
• Tackling cargo demand specific challenges: imbalanced data, skewed data distributions, features engineering
• Performing in-depth model evaluation to properly assess live model performances and identify data drift or areas for improvements
• Taking part in the improvement of our training pipeline framework (tests, automation..)
• Survey state of the art models / practices in our domain specific applications

• Backend: Python 3.11+ with SQLAlchemy
• Orchestration: Argo workflows over an auto-scaled Kubernetes cluster
• Datastores: Postgresql and Druid
• Common ML libraries/tools: LightGBM, XGBooost, TensorFlow/Keras, Pandas, Dask, Dash, Jupyter notebooks, SHAP
• Model versioning and registry tool: MLFlow
• Gitlab / Kubernetes for CI/CD
• Prometheus/Grafana and Kibana for operations

• You are currently pursuing a Master's Degree in data science, computer science or applied mathematics
• You have a good understanding of fundamental machine learning concepts, including classification and regression tasks, as well as their underlying mathematical principles
• You have a good understanding of basic evaluation metrics : accuracy, recall, precision, RMSE etc, as well as basic statistics
• Hands-on experience with python language and standard data tools: Pandas, NumPy
• You have a strong curiosity and desire to learn
• You have a solid computer science background in Python
• You have a good experience in model development (Kaggle, class or personal projects,..)

By joining us, you will integrate:

• A self-financed startup with a strong technical DNA
• Beautiful 800 m² offices in the heart of Paris (Bd Poissonnière)
• Attractive remuneration
• A caring and stimulating team that encourages skills development through initiative and autonomy
• A learning environment with concrete opportunities for evolution ‍

You will also benefit from:

• Training on demand
• A hybrid policy: a day of remote work per week and the possibility to work occasionally from abroad
• A great company culture (monthly afterworks, regular meetings on technology and products, annual off-site seminars, team-building…)

• A screening with the manager

• A technical test to prepare at home

• A technical interview at our offices with the manager

Wiremind is committed to equality of opportunity, diversity, and fairness. We encourage all candidates with the necessary experience to apply for our job offers.