AI Research Apprentice

About the role

Origin (formerly 10xConstruction), which builds autonomous robots for US construction sites, is hiring an AI Research Apprentice in Bengaluru. You will prototype diffusion-based vision models, vision-… Subscribe to Premium to view the full description and apply.

Responsibilities

• Design and train diffusion-based generative models for high-resolution synthetic data.
• Train and improve segmentation and surface-quality models used on robots in the field.
• Develop training objectives for vision-language-action models that link work orders to segmentation masks and robot actions.
• Build large-scale auto-annotation pipelines for images and point clouds.
• Benchmark accuracy, latency and memory on Jetson-class hardware and compress models with distillation or LoRA.
• Integrate research prototypes into live ROS 2 stacks with the perception and robotics teams.

Requirements

• Strong foundation in deep learning, probabilistic modelling and computer vision through coursework or research.
• Hands-on experience with diffusion models in PyTorch or JAX.
• Familiarity with multimodal transformers and VLMs such as CLIP, BLIP or LLaVA.
• Working knowledge of active learning, self-training and pseudo-labelling.
• Solid Python and PyTorch skills with Git workflows; C++ and CUDA are a bonus.

Benefits

• Work on AI for robots deployed on live construction sites.
• Join a robotics team from IITs, Stanford and UCLA.

Required Skills