Honda Research Institute (HRI) in Mountain View, CA- MS or PHD

Honda Research Institute (HRI) in Mountain View, CA, has multiple research internship openings for PhD (and qualified MS) students.
Summer 2017 internship positions:

Accelerated Deep Learning for Computer Vision (Job Number: P16INT-01)
Image/Video Captioning for Information Retrieval (Job Number: P16INT-02)
Motion Planning for Autonomous Driving (Job Number: P16INT-03)
Multimodal Data Segmentation (Job Number: P16INT-04)
Sequence Anomaly Detection (Job Number: P16INT-05)
Reinforcement Learning for Autonomous Navigation (Job Number: P16INT-06)
Traffic Modeling and Simulation (Job Number: P16INT-07)
Visual Localization/SLAM (Job Number: P16INT-08)
Visual Attention and Dynamic Scene Saliency (Job Number: P16INT-09)
Visual 3D Object Detection and Pose Estimation (Job Number: P16INT-10)

Please see below for the detailed descriptions of each position.

——————————————————————–

How to Apply: Please send an e-mail to interns@honda-ri.com with the following:
·        Subject line including the job number(s) you are applying for.
·        Recent CV
Candidates must have the legal right to work in the U.S.A.

——————————————————————–

Accelerated Deep Learning for Computer Vision (Job Number: P16INT-01)

This position focuses on developing novel computer vision techniques for real-time object detection. There is a strong focus on using deep neural networks and making them run faster. The candidate’s responsibility includes the analysis of the network architecture, developing generalizable techniques to speed up the training and inference of Deep Neural Networks without sacrificing output quality, and implementation on an automotive platform.

Qualifications:
·        Research experience on Machine Learning and Computer Vision.
·        Prior experience with Caffe is a plus.
·        Excellent programming skills in C++ and/or Python.

——————————————————————–

Image/Video Captioning for Information Retrieval (Job Number: P16INT-02)

This position focuses on research and prototyping on image/video description, and scene understanding system. The successful candidate’s responsibilities include designing novel algorithm for scene description/understanding systems to achieve intelligent search function from multi-modal corpus, including video and sensor data. Strong background in Machine Learning is preferred.

Qualifications:
·        Research experience on Natural language processing, Computer vision or Information retrieval.
·        Experience designing deep convolutional neural networks using Caffe, TensorFlow, Theano, etc..

  •        Excellent programming skills in C++ and/or Python.
    ·        Hands on experience multi-modal sensor data is a plus.

    ——————————————————————–

    Motion Planning for Autonomous Driving (Job Number: P16INT-03)

    The job function includes active contribution to HRI activities related to self-driving technology. Responsibilities include: Theoretical analysis, development of motion planning algorithms, software development in C++/ROS/Linux and integration of the developed ROS modules with the existing modules at HRI. Background related to the following areas is required: motion planning in the presence of dynamic obstacles, real time trajectory generation, path optimization and/or collision avoidance algorithms.

    Qualifications:
    ·        Demonstrated programming skills in C++
    ·        Hands-on experience with ROS
    ·        Comfortable in development in Linux

    ——————————————————————–

    Multimodal Data Segmentation (Job Number: P16INT-04)

This position focuses on research on data management for development of next generation mobility systems. Responsibilities include design and implementation of segmentation/clustering algorithms for multimodal/multisensory data, and development of an experiment protocol to benchmark the proposed algorithm. A strong background in computer vision, machine learning, video analytics or signal processing is preferred.

Qualifications:
·        Strong familiarity with machine learning, e.g., unsupervised Bayesian machine learning
·        Experience designing deep convolutional neural networks using Caffe, TensorFlow, Theano, etc..
·        Excellent programming skills in C++ or Python
·        Hands on experience on acquisition of multi-modal sensor data is a plus

——————————————————————–

Sequence Anomaly detection (Job Number: P16INT-05)

This position studies driver / vehicle / road scene anomaly detection using multi-model sensory signals including road camera, driver camera, driver physiological sensors, Lidar scanner, etc. The intern will collaborate with the researchers at HRI to design advanced machine learning framework. A strong familiarity with anomaly detection, sequence modeling and deep learning is preferred.

Qualifications:
·        Excellent programming skills in C++ or Python.
·        Research experience in human machine interaction and related machine learning methods.
·        Strong publication record in the areas of, anomaly detection, human-machine interaction.
·        Hands on experience with sensing devices. Experience in ROS is a plus.

——————————————————————–

Reinforcement Learning for Autonomous Navigation (Job Number: P16INT-06)

This position focuses on applying reinforcement learning methods to autonomous navigation problems. Application topics include learning to follow a target (vehicle, person) in dynamic environments or multi-agent collaborative navigation. Responsibilities include: surveying latest RL research, algorithm development, software development in simulation, and comparison with state-of-the art algorithms. Background related to the following areas is required: machine learning, reinforcement learning, policy optimization.

Qualifications:
·        Excellent programming skills in Python
·        Demonstrated experience in machine learning, preferably deep reinforcement learning
·        Practical experience in Deep Learning platforms such as OpenAI Gym or Tensorflow is a plus

——————————————————————–

Traffic Modeling and Simulation (Job Number: P16INT-07)

This position focuses on development of frameworks for scenario creation using real-world driving data and implementation of intelligent traffic participant models in a simulation environment. The candidate’s responsibilities include: software development of complex simulation models in a 3D world, survey and implementation of existing traffic participant models and comparison of simulated behaviors to real-world data, deployment cloud-based simulations.  A strong background in 3D modelled traffic simulation and participant behavior modelling is preferred.

Qualifications:
•        Excellent programming skills in C++
•        Strong experience with ROS/Gazebo
•        Demonstrated experience in data-driven behavioral modelling a plus

——————————————————————–

Visual Localization/SLAM (Job Number: P16INT-08)

The candidate’s responsibilities include: designing novel algorithms that enable robust visual localization/SLAM under challenging situations (e.g. varying lighting conditions), data processing from camera, IMU, LiDAR and other sensors, development of the software, and time permitting, implementation on the vehicle and evaluation the real-time performance of the developed technologies. A strong background in robotics, computer vision, and machine learning and research experience in visual localization or SLAM is preferred.

Qualifications:
·        Excellent programming skills in C++ and python/MATLAB under Linux
·        Experience in deep learning and visual inertial navigation is a plus
·        Publication record in top conferences (ICRA/RSS/ICCV/CVPR, etc..) is a plus

——————————————————————–

Visual Attention and Dynamic Scene Saliency (Job Number: P16INT-09)

This position focuses on analysis of driver’s visual attention from scene and driver’s behavior information utilizing driving data consists of vision-, range-  as well as vehicle-state and driver monitoring data for next generation advanced driver assistance systems. The candidate’s responsibilities include: designing novel algorithm using computer vision and machine learning approaches to understand driver’s visual attention and scene saliency, software development for evaluation for the proposed algorithms. Strong background in human-machine interaction, machine learning and Bayesian inference is preferred.

Qualifications:
·        Research experience in Computer vision, Lidar/Radar point-cloud data processing.
·        Experience designing deep neural networks using Caffe, TensorFlow, Theano or similar tools.
·        Excellent programming skills in C++ or Python.
·        Hands on experience with sensing devices such as eye tracker and experience in ROS is a plus.

——————————————————————–

Visual 3D Object Detection and Pose Estimation (Job Number: P16INT-10)

This position focuses on 3D object detection and orientation estimation from a single monocular image for automated driving and robotics applications. Specifically, the work content includes design and implementation of an algorithm to generate 3D bounding boxes for objects and their orientation from a given image. The candidate is responsible for development of an experiment protocol to benchmark the proposed algorithm. A strong background in computer vision, 3D object detection/tracking and/or machine learning is preferred.

Qualifications:
·        Strong publication record in the areas of computer vision and machine learning
·        Experience designing deep convolutional neural networks using Caffe, TensorFlow, Theano, etc..
·     Excellent programming skills in C++ or Python.