Staff ML Engineer leading Reddit’s ads retrieval modeling: designing candidate-generation and two-tower retrieval systems, running experiments, and setting technical direction while mentoring engineers on the Retrieval & Targeting team.
Principal Data Scientist for Reddit's User Ecosystem, setting the scientific direction for consumer problems such as discovery, personalization, growth, retention, and communities. Day to day: define measurement and experimentation strategy, build user/content/community models and causal frameworks at scale, and guide product decisions. Core technologies: machine learning, deep learning (PyTorch o
Leads the Core Ads Ranking team at Reddit, defining technical vision for foundational ML models, user understanding, and graph learning; oversees end-to-end model lifecycle from research to production using TensorFlow/PyTorch, and collaborates with cross-functional teams to improve ad ranking systems.
Staff Machine Learning Engineer building advanced ML models (DNNs, transformers) for Reddit's Ads Ranking system, focusing on app ads modeling, feature engineering, and large-scale recommendation systems using TensorFlow/PyTorch.
Staff ML Engineer leading Reddit's Ads Retrieval team, designing large-scale candidate generation models (two-tower, embeddings, ANN) using TensorFlow/PyTorch to improve ad relevance for 130M daily users.
Senior Data Scientist on Reddit's Ads team building ML models and statistical solutions to optimize advertiser experience and platform performance, using Python, Spark, and large-scale data analysis.
Senior Software Engineer in Motorola Solutions' camera systems group, designing and optimizing modern C++ for embedded systems and deploying AI/computer vision models on camera hardware for real-time safety and security video analysis.
A two-year graduate programme in Bristol where you’ll build software connecting Graphcore’s ML accelerators with PyTorch and Triton, implementing features, optimising performance, fixing defects, and writing tests and documentation while learning about compilers and runtime software for AI workloads.
A graduate software engineer tests AI and machine learning workloads on Graphcore accelerator hardware, creating ML workloads, analysing numerical accuracy and performance, designing tests, and building CI/CD pipelines using Python, C++, Linux, and ML frameworks such as PyTorch, JAX, Triton, or TensorFlow.
We are looking for a Senior Solutions Architect, Industrial Robotics to engage with customers, partners, and other NVIDIA software engineers to support applications and platforms that leverage NVIDIA technology that…
NVIDIA has been transforming computer graphics, PC gaming, and accelerated computing for more than 25 years. It’s a unique legacy of innovation that’s fueled by great technology—and amazing people. Today, we’re tapping…
Lumen is the trusted network for the AI‑powered world, connecting people, data, and applications through our expansive fiber network and connected ecosystem. We enable secure, high‑performance connectivity across…
Designs and optimizes high-performance CUDA kernels and GPU libraries for AI training, inference, and HPC workloads. Works with ML engineers to profile production GPU systems using Nsight and NCCL, shipping performance improvements in C++/CUDA, PyTorch, JAX, or Triton.
Designs, builds, and operates the GPU/accelerator platform layer for large-scale AI training and inference, including scheduling, storage, networking, and developer tooling for ML engineers. Uses Python, Go/C++, Kubernetes, Slurm/Ray, and cloud ML infrastructure; 100% remote U.S. role.
This position does not provide sponsorship for employment now, or in the future, in any format. Position Summary: The Applied Artificial Intelligence Engineer is part of the Education Design and Informatics team within…
Member of Technical Staff for Model Training infrastructure and platform at Xperteez Technology. The role involves architecting and scaling distributed systems for supervised fine-tuning, reinforcement learning, and distillation across GPU clusters, owning inference optimization, experiment tracking, and platform architecture for AI research teams, using Python and systems languages such as C++, R
Job Title: Member of Technical Staff, Model Training (applied research) Job Type: Full-time Location: Remote The Role In this role, you will work across language models, multimodal systems, robotics, and autonomous…
Job Title: Member of Technical Staff, Model Training (infra & platform) Job Type: Full-time Location: Remote The Role In this role, you will take technical ownership of the distributed systems and platforms that…
Researches and scales advanced post-training methods (SFT, RL, distillation) for large language and multimodal models, designing experiments to improve reasoning, generalization, and agent capabilities using Python and PyTorch.
Cresta unlocks the true potential of the customer experience, turning every conversation into a competitive advantage. Cresta’s unified AI platform combines conversational AI agents, real-time human agent augmentation,…
Cookies?
A session cookie keeps you signed in. If you accept, we also use analytics cookies (Google Analytics and PostHog) to understand usage. Learn more
We couldn't check your fit for this role — add a CV to your profile to see it next time.