Machine Learning Engineer - Foundational

Harmattan AI

ABOUT US Harmattan AI is a next-generation defense prime building autonomous and scalable defense systems. Following the close of a $200M Series B, valuing the company at $1.4 billion, we are expanding our teams and capabilities to deliver mission-critical systems to allied forces. Our work is guided by clear values: building technologies with real-world impact, pursuing excellence in everything we do, setting ambitious goals, and taking on the hardest technical challenges. We operate in a demanding environment where rigor, ownership, and execution are expected. About the Role We are looking for a Machine Learning Engineer to join our Semantic Scene Understanding team in Paris. In this role, you will design the core algorithms to extract semantic information in real-time from the theatre of operations as seen through the different cameras of our different UAVs, to improve the operator’s scene understanding. Responsibilities * Design and Train: Develop state-of-the-art machine learning algorithms for semantic segmentation, object detection, and classification tailored to aerial imagery. * Advanced Feature Extraction: Build high-level tactical features on top of base semantic data, such as real-time road vectorization, trafficability analysis, and dynamic obstacle mapping. * Multi-Agent Fusion: Architect pipelines that temporally and spatially align semantic data from multiple moving UAVs into a cohesive Common Operational Picture (COP). * Edge Optimization: Optimize and deploy these algorithms directly into our tactical C2 platform, utilizing quantization, pruning, and hardware acceleration to meet strict real-time compute constraints. Candidate Requirements * Educational Background: MSc in Computer Science, Machine Learning, or a related field. A PhD is a strong plus. * Foundational Knowledge: Deep understanding of Machine Learning theory, Linear Algebra, and 3D-Geometry algorithms. * Core Tech Stack: Expert-level command of Python and deep learning frameworks (PyTorch). * Performance Engineering: Experience with C++ and inference optimization frameworks (e.g., TensorRT, ONNX Runtime, CUDA) is highly desirable. * Domain Experience (Plus): A track record of shipping CV/ML algorithms in production, particularly for edge/embedded systems or involving aerial (EO/IR) imagery. * Strong Ownership: Ability to take a feature from an ArXiv paper all the way to a ruggedized tactical PC. * Adaptability & Mission Focus: Thrives in a fast-paced startup environment and is 100% dedicated to building ethical defense technologies that bring a strategic edge to allied nations. Communication: Excellent verbal and written communication skills to collaborate effectively with software engineers and hardware teams. We look forward to hearing how you can help shape the future of autonomous defense systems at Harmattan AI.

Vestiaire Collective

Vestiaire Collective is the leading global platform for desirable pre-loved fashion and a pioneer in transforming how people consume fashion.   Our mission is simple: make circular fashion the norm, not the exception. Through technology, expertise, and a highly engaged global community, we enable millions of people to buy and sell fashion in a more sustainable way.   Founded in Paris in 2009, Vestiaire Collective is now a globally scaled marketplace with offices in Paris, London, Berlin, New York, Singapore, and Ho Chi Minh City, and logistics hubs across Europe, Asia, and the US. Today, we are a team of around 600 people from over 50 nationalities, united by a shared ambition: to drive meaningful change in the fashion industry.   Our values, Activism, Transparency, Dedication, Greatness, and Collective, shape how we build, collaborate, and grow every day. About the Role We are seeking a Foundational Machine Learning Engineer for a high-impact greenfield opportunity to build our MLOps infrastructure from the ground up at Vestiaire Collective. While driving our AI authentication initiatives (deploying multi-model approaches including computer vision for luxury product authentication and counterfeit detection) will be your immediate focus, your long-term mission will be to scale foundational architecture across the entire marketplace. You will expand our ML capabilities to power broader domains, primarily focusing on search and recommendation systems, with future expansions into dynamic pricing and marketing technologies. Acting as the bridge among Applied Science, Data Platform, and Backend Engineering, you will design robust, decoupled architectures and spearhead the MLOps strategy with our Director of Data, prioritizing system maintainability, engineering hygiene, and the reliable deployment of complex models, ensuring all our ML models across the board deliver high-throughput, low-latency business impact. What You Will Do Short-Term Impact (First 6 Months): Partner closely with the Operations squads and Data Scientists to accelerate ML and RAG prototypes into resilient, production-ready code. You will directly integrate with the team to deploy, optimize, and scale heavy-width CV and VLM models focused on fraud detection and luxury product authentication, immediately improving our trust and safety ecosystem. Mid-Term Foundation (MLOps Lifecycle & Infrastructure): Lead the end-to-end foundational groundwork of our ML lifecycle by designing robust systems for Data & Feature Management, Model Tracking & Registry, and Model Serving & Monitoring. You will scale infrastructure by automating continuous retraining pipelines that handle diverse deployment cadences (from daily fraud detection to weekly recommendations), design resilient multi-model architectures, and critically evaluate the technical overhead and TCO of our in-house tools against enterprise-grade platforms to ensure long-term resilience. Long-Term Vision (Centralizing 360-Degree MLE Capabilities): Act as a pioneer and cornerstone hire for the ML engineering discipline at Vestiaire Collective, setting the technical standards to help scale the AI/ML organization. You will transition into a centralized foundational role, moving beyond single-squad operations to mentor the team and provide horizontal ML infrastructure support to multiple domains, including Search, Discovery, Pricing, Marketing, and Data Platforms. Who You Are Must-Haves: Experience: 5-8+ years of hands-on experience in Machine Learning Engineering, specifically focused on building and scaling MLOps infrastructure and productionizing ML systems. Production Infrastructure: Proven expertise in deploying low-latency, high-throughput ML inference services (using FastAPI, TorchServe, Triton Inference Server, or Ray Serve) across both classical lightweight and heavy-width ML models (PyTorch/TensorFlow). Strong preference for AWS (EKS, EC2, SageMaker) / Snowflake and Open Source ecosystems over GCP/Azure. MLOps & Pipelines: Deep experience building automated, continuous model retraining pipelines to handle concept drift (ranging from daily to weekly cycles). You have orchestrated decoupled, multi-model AI architectures using tools like Airflow, Kubeflow, or Metaflow, and possess strong expertise in model registry and tracking tools like MLflow or Weights & Biases. Feature Stores: Hands-on experience evaluating, building, or extensively leveraging online (Redis, DynamoDB) and offline (Snowflake, S3) Feature Stores in a production environment. Familiarity with frameworks like Feast or custom dbt-based pipelines is highly valued. Strategic Builder Mindset: You are an analytical builder who thinks long-term. You can successfully evaluate TCO for bespoke internal systems versus enterprise tools, anticipate technical liabilities, and design robust architectures that handle unpredictable peak traffic surges. Collaboration & Engineering Hygiene: Strong cross-functional communication skills. You excel at translating complex ML prototypes into highly scalable production code backed by strict version control, rigorous testing, and CI/CD best practices, seamlessly connecting data science innovation with backend engineering execution. Nice-to-Haves: Relevant Domain Expertise: Background in E-commerce, Single-SKU Marketplaces, Search & Recommendation, Trust & Safety, or Counterfeit Detection. Vision, Edge & Optimization: Hands-on experience with Vector Databases, Visual RAG pipelines, deploying Deep Learning VLM models, and optimizing models for edge computing or low-latency inference (e.g., ONNX, TensorRT). Infrastructure & Observability: Advanced experience with containerization (Docker, Kubernetes), Infrastructure as Code (Terraform), and data transformation workflows (dbt). Familiarity with setting up advanced monitoring for model performance, concept drift, and system health (Datadog, Prometheus).

Voodoo

ABOUT VOODOO Founded in 2013, Voodoo is a tech company that creates mobile games and apps with a mission to entertain the world. Gathering 800 employees, 7 billion downloads, and over 200 million active users, Voodoo is the #3 mobile publisher worldwide in terms of downloads after Google and Meta. Our portfolio includes chart-topping games like Mob Control and Block Jam, alongside popular apps such as BeReal and Wizz. TEAM The Engineering & Data team builds innovative tech products and platforms to support the impressive growth of their gaming and consumer apps which allow Voodoo to stay at the forefront of the mobile gaming industry. The Voodoo Ad-Network is an autonomous product group of around 60 highly driven professionals with an ambitious mission: building top-tier ad network services. Our primary goal is to leverage Voodoo’s massive first-party data ecosystem to optimize and scale monetization. We are in a rapid growth phase, expanding into new ventures such as opening to external inventory, penetrating the external advertiser market, and driving social network monetization following our recent acquisition of BeReal. To support this incredible trajectory and promising early results, we are scaling our team. The Feature Platform Team is the foundational infrastructure engine empowering our ML Ads Recommendation capabilities. Drawing inspiration from industry-leading feature stores, we build and maintain the unified data layer for our machine learning features. Our mission is to accelerate the ML lifecycle by providing a unified, scalable, and highly available architecture for computing, storing, and serving batch, real-time, and on-demand features. Beyond just building the infrastructure, we are a highly proactive team that continuously explores new data signals and feature engineering opportunities to push the boundaries of our targeting performance. This role is a hybrid position, either based in Helsinki, Paris or Strasbourg. ROLE We’re looking for a Senior Data Engineer to join our Feature Platform Team. You will be joining a dedicated squad of Data and ML Engineers focused on guaranteeing consistency across offline training and online inference, eliminating training-serving skew, and enabling our Data Scientists to seamlessly and rapidly deploy the next generation of high-performing models. In this role as a Senior Data Engineer, your scope extends far beyond classical data engineering. You will be responsible for managing both the offline and online components of our machine learning architecture. This means bridging massive-scale data processing (handling both heavy batch jobs and subsecond real-time feature updates) with high-load online services that must process and return features for inference with strict low latency. * Architectural Ownership: Take end-to-end ownership of highly visible projects from initial ideation to production release. This includes feature scoping, timeline estimation, architecture design, and benchmarking next-generation technologies. * Proactive Data Innovation: Go beyond passive implementation by actively partnering across the entire data lifecycle. You will deeply understand the downstream Data Science domain to discover high-impact feature opportunities, while also collaborating closely with upstream data engineering teams to understand ingestion mechanisms (right up to the SDK and Bidding platforms) to unlock and integrate new data signals. * ML Infrastructure & Feature Platform: Collaborate closely with Data Scientists and ML Engineers to design, scale, and optimize core components spanning both offline training and online inference, including our Feature Store (supporting batch, streaming, and low-latency on-demand computation) and engines for on-demand training dataset generation. * Pipeline Engineering: Build, maintain, and optimize mission-critical data pipelines spanning both extensive batch processing and continuous real-time streams (ensuring subsecond feature updates) to adapt to ever-evolving business and machine learning needs. * High-Performance Online Services: Actively build and maintain the high-load backend applications that power our ML model serving, ensuring they can process and return features with low latency and high availability under heavy traffic. * Scalability & Performance: Work hand-in-hand with our infrastructure teams to guarantee the reliability, security, and immense scalability required for an ad-network ecosystem. * Agile Collaboration: Thrive in a fast-paced agile environment with rapid decision-making processes. You will collaborate daily with back-end developers, data scientists and product managers. * Mentorship & Team Culture: You will actively contribute to our engineering culture, share knowledge, and ensure every team member feels comfortable, supported, and empowered to grow in their role. PROFILE We are looking for a Senior Data Engineer who deeply understands both the data lifecycle and the specific challenges of putting machine learning models into production at scale. * 6+ years of proven experience as a Data Engineer, ML Engineer, Backend Engineer, or a closely related role in a high-scale environment. * Big Data & Streaming Mastery: Extensive hands-on experience working with Flink or Spark at scale. Deep expertise in Flink (or similar stateful streaming platforms) as you will be a key contributor in scaling our real-time streaming architecture. * Coding Proficiency: Advanced expertise in Python for robust ETL pipelines and custom feature-definition SDKs/DSLs * Experience or a strong willingness to work with Golang for building high-performance, low-latency backend applications as well as familiarity with Java is highly valued for our Flink streaming workloads. * Data Architecture: Deep understanding of modern Data Lakehouse design principles, open table formats (like Iceberg), optimization techniques, and data modeling. * Cloud & DevOps: Strong hands-on experience with a major cloud platform (AWS, GCP, Azure, etc.), though AWS is our preferred environment. Familiarity with DBT for data pipeline transformation is a plus. * ML Production Awareness: You have a solid grasp of the unique challenges involved in running ML models in production, including working with Feature Stores, mitigating training-serving skew, and model monitoring. * System Design: You are highly familiar with topics surrounding system scalability, high availability/reliability, low-latency API design, and security best practices. OUR STACK * Languages: Python (ETL & SDKs), Golang (High-Performance Online Services), Java (Flink) * Processing & Orchestration: Spark, Flink (Real-time Streaming), Airflow, DBT * Storage & Infrastructure: Apache Iceberg, Amazon Web Services (AWS), Kubernetes, Terraform BENEFITS * Competitive salary based on experience * Swile Lunch voucher * Gymlib (100% covered by Voodoo) * Premium healthcare coverage with SideCare, 100% covered for you and your family * Wellness activities in our Paris office * Remote Fridays