Aprendizaje automático, Snap Inc.

Our engineers utilize state of the art models and continually push the boundaries of what’s possible in Snapchat's content, monetization, infrastructure, features, and more!

AI Lenses

Unified Recommendations

Dynamic Ads

Machine Learning Engineering at Snap

As a Machine Learning Engineer at Snap you’ll drive Snapchat’s dynamic experience through the full lifecycle of advanced state-of-the-art models - from data preprocessing, feature engineering, and model training, to deployment and ongoing optimizations. You’ll leverage cutting-edge techniques, ranking algorithms for ad relevance, recommendation engines for personalized content and NLP for enhanced interactions - all while processing petabytes of data for over 850 million users.

Through both classic and deep learning models, you’ll create precise, responsive experiences that empower users to express themselves, connect and discover the world in real-time.

“The real benefit that Snap has is the size of our scale, and the breadth of influence and impact that people will have. You can run fast, have broad influence and actually see your work hit production with the right experimentation tools and infrastructure to be productive.”

Equipos en aprendizaje automático

Our machine learning engineers solve real world ML problems.

Monetización

Como ingeniero de aprendizaje automático en el equipo de Monetización, construirás y optimizarás todo el ecosistema de anuncios. Impulsarás una alta relevancia y un alto impacto no solo para los anunciantes y los usuarios, sino para todo Snap. Desde el diseño de sistemas de alto rendimiento para pujas en tiempo real o la publicación de anuncios y subastas, la personalización de clasificadores ligeros y pesados, la creación de soluciones para la orientación y la entrega de anuncios, continuarás garantizando una integración perfecta de los anuncios en toda la plataforma. Entrenarás modelos con miles de millones de ejemplos, usando aprendizaje multitarea, modelado de secuencias y modelado de interacción de usuario x anuncio. Nuestros modelos predicen la información demográfica del usuario para mejorar la orientación al público con redes neuronales y contenido de gráficos, para comprender cómo nuestro trabajo da forma al futuro de la plataforma de anuncios de Snapchat.

En qué trabajarás:

Publicidad orientada por IA
Nuevos productos y experiencias de anuncios personalizados
Propietario del principal impulsor de ingresos de Snap
Desarrollo de productos publicitarios de vanguardia

Contenido

Como ingeniero de aprendizaje automático en el equipo de Contenido, trabajarás en el desarrollo de funciones para modelos de recomendación de video de formato corto altamente visibles que impulsan la creatividad y la conexión en el centro de nuestra plataforma, como Historias y Spotlight. Te centrarás en mejorar la experiencia del usuario, impulsar la innovación en Spotlight, la clasificación de los Lentes, apoyar a los creadores de contenido y lograr que la creación de contenido no requiera esfuerzo. Nuestro equipo aborda la comprensión de contenido multimodal, utilizando el aprendizaje avanzado de representación de gráficos y la clasificación de contenido para ofrecer las mejores experiencias para nuestros usuarios. Optimizamos los modelos de aprendizaje automático por medio de la revisión humana y aprovechamos técnicas de personalización de última generación, lo que nos ubica a la vanguardia en la innovación de contenido de video de formato corto.

En qué trabajarás:

Recomendaciones de contenido personalizadas
IA generativa para creación de contenido
Desarrollo de funciones para productos de alta visibilidad como Spotlight, Historias, el feed Descubrí y las sugerencias de amigos

Core Growth

As a Machine Learning Engineer on the Core Growth team you’ll drive the search and friending experience across 15+ products, supporting over 100 million daily active users. Our team powers search for friends and content using advanced technologies, including graph databases, Large Language Models (LLMs), Mixture of Experts (MoE), and sophisticated machine learning stacks. You’ll develop search algorithms that connect users through friend and communication graphs, use visual and text embeddings for content discovery and enhance search with pretype recommendations that predict users queries. In addition to optimizing search infrastructure, the team’s friending products help users make real, meaningful connections to responsibly grow the Snapchat community. This work strengthens user engagement and retention, enabling other Snapchat features like messaging and story maps to flourish at scale.

What you’ll work on:

Create AR experiences
Large scale system (5B notifications daily)
Snapchat’s graph with 1+billion of nodes and hundreds of billions of edges

Platform

As a Machine Learning Engineer on the Platform Engineering team you’ll build models that provide timely, accurate, high quality and actionable data for decision making that forms the backbone of Snap's external reporting, experimentation, and ML/Ranking. Snap’s AI platform (aka Bento) powers many large scale ranking, recommendation, and Gen AI use cases. Over the past year, our scale of operations increased by 4-10X. You’ll develop embeddings that are consumed by downstream modeling teams. This team is mission critical to scaling and diversifying Snaps ML model offerings.

What you’ll work on:

Highly available, multi-cloud infrastructure
Optimizing the entire developer experience
Ship high-quality features fast

Trust & Safety

We want Snapchat to be a safe and positive experience for everyone who uses our platform or products. For this reason, we created rules and policies that explain the rights and responsibilities of all members of our community.

What you’ll work on:

Build Privacy-preserving AI
Launch new safety features

Generative AI

Our Generative AI team is at the forefront of innovation, shaping the future of augmented reality (AR) through SnapAR, which includes Lens Studio, Lens Engine, and Lens Cloud. Our engineers build real-time advanced AI models and machine learning models for dynamic transformations, such as realistic filters, object recognition, and scene understanding, ensuring seamless AR interactions for products like ML Lenses and Dreams. Lens Studio 5.0 stands out as the leading developer platform for creating cutting-edge Lenses that blend generative AI with traditional computer vision techniques. Generative AI empowers creators to design engaging lenses effortlessly without requiring expertise in 3D modeling, animation or coding.

What you’ll work on:

Develop and optimize Generative AI Models
Build AI-driven tools like text-to-3D generation
Deliver generative AI and computer vision powered features

Locations

Our RTO (Return to Office) policy is 4 times per week hiring in these office locations

Aprendizaje automático en Snap

Escuchá al equipo sobre la vida en Snap en aprendizaje automático

We're Hiring!

Our interview process covers engineering, foundational, and applied ML.

Coding

Expect to solve algorithmic problems that test your proficiency in data structures, algorithms, and problem-solving skills. Focus on your ability to write clean, efficient, and well-documented code.

ML Fundamentals

You’ll be assessed on ML theory and core machine learning models, concepts, techniques and applications. Be prepared to discuss supervised and unsupervised learning, recommendation systems, ranking, model evaluation metrics, and optimization techniques.

ML Applied Design

Evaluates your ability to design and apply machine learning solutions to real-world problems. You may be asked to walk through the end-to-end process of selecting models, feature engineering, and evaluating performance. At times this can test your ability to problem solve in an ambiguous environment.

ML System Design

The focus will be on designing scalable and robust ML systems that can handle large-scale data and production environments. Expect to discuss the infrastructure and trade offs in architecture, model deployment strategies and system monitoring.

Our Interview Process