Machine Learning Engineer Intern

At Datature, we are building a platform for training and deploying machine learning models swiftly without a single line of code. Datature Nexus is an MLOps platform that allows users to label unstructured data and construct complex machine learning pipelines all in their browsers using highly visual editors. To date, Datature helps more than 10,000 teams worldwide build computer vision solutions in niched, complex domains such as medicine, agriculture, and manufacturing with our end-to-end data platform. The Scope 🔭 As part of the integration team, you will be working directly with customers to integrate our MLOps Platform directly into their stack, and ideate ways to optimize their computer vision implementation. You will be expanding our neural network offerings, cutting-edge features, and model observability functions for our users. In addition, you will be working on Datature’s Insights and MLOps offering to help users build better, fairer, models. In addition, you will scout and prototype state-of-the-art vision algorithms and methods to provide our users and the team with potentially disruptive technologies that can be integrated into the Datature platform to improve user value. Ideally, you enjoy research, prototyping, and sharing new ideas based on real-user feedback and problems to work on a roadmap of ambitious features that we ship to our users in our quarterly platform updates. Requirements ⚒ Some experience using PyTorch, JAX, or Hugging Face Transformers Fluent in Python and at least one system programming language (C, C++, Rust) Interest in Computer Vision, Vision-Language Models, and Scene Understanding Algorithms Exposure to working with Relational and No-SQL Databases Some knowledge of Google Cloud Platform and Linux/Bash Knowledge in Parameter-Efficient Fine-Tuning (LoRA, QLoRA), Quantization (INT8, FP8, AWQ), and Open-Vocabulary Detection Exposure to modern VLM architectures - vision encoders (SigLIP, DINOv2), projection layers, autoregressive decoders Exposure to building your own Instance Segmentation / Object Detection / Grounding Models Exposure to inference serving and optimization (vLLM, SGLang, TensorRT, ONNX Runtime) Interest in Vision-Language-Action Models, embodied agents, and robot perception About You 🤖 You don't need a Computer Science degree to write great software. We are looking for hackers and enthusiasts who have participated in hackathons / competitive programming / open-source development. Your coding projects outside of school curriculums matter more than your grades - we'd love to discuss your most passionate projects rather than a list of curriculum-mandated projects. Why You'll Love Us 🎈 We are a technology-first team dedicated to developing the most intuitive and high-performance platform to transform how users build machine learning capabilities. Our team is a quirky group of engineers passionate about discussing code and algorithms, and you'll find yourself in stimulating debates that spark innovation. At Datature, you'll tackle critical challenges at the intersection of machine learning, data processing, and client engineering, making a significant impact on the model architectures and optimization solutions available to our users. While building development tools for a single company is rewarding, joining Datature means contributing to tools that will empower the next 10,000 companies. If this sounds like your cup of tea, we’d love to hear from you.