Robot Learning Engineer

Haptica Robotics; San Francisco (CA); On-site; USD 150,000 - 210,000; Full time

Job summary

Haptica Robotics is seeking a Robot Learning Engineer to own the end-to-end learning function in our San Francisco office. You will build the training pipeline, stress-test sensors, and evaluate policies that leverage tactile data. You’ll collaborate with hardware, data ops, and customers to integrate data into diverse stacks and demonstrate impact through experiments and publications.

The role emphasizes autonomy, strong ML background, and hands-on hardware experience with cameras and sensors.

Qualifications

• MS or PhD in robotics, CS, EE, or related field, or equivalent industry experience.
• Extensive production-level software and ML engineering practice.
• Experience with PyTorch and modern DL frameworks.
• Strong written and verbal communication; ability to simplify complex topics.
• Experience running imitation learning from data collection to training and debugging.
• Comfort with hardware: cameras, IMUs, calibration, and sensor time synchronization.

Responsibilities

• Build data collection, training, and inference pipelines on standard hardware with Haptica.
• Train imitation learning policies (ACT or diffusion-based) using tactile data for key tasks.
• Define evaluation benchmarks and reuse public setups to quantify data impact.
• Extend data collection to egocentric glove data, integrate hand tracking and retargeting.
• Inform hardware decisions on sampling rate, resolution, synchronization, and calibration.
• Share findings via demo videos, writeups, and papers with academic collaborators.

Skills

Robotics

ML engineering

PyTorch

Imitation learning

Hardware familiarity

Communication

Education

MS or PhD in robotics, CS or related field

Tools

PyTorch

Job description

ABOUT HAPTICA

Haptica is building the standard for tactile in robotics.

We build pneumatic-based tactile sensing wearables for robots and people. Gloves and sensors that integrate across grippers, human hands, and robot hands. Teleop teams use Haptica to give human operators low-latency haptic feedback that replicates the pressure that robots put on an object. Data ops teams use Haptica to collect UMI, egocentric, and telop data to train their models.

Haptica was founded by Dr. Cosima du Pasquier and Paola Peraza Calderon in early 2026 and is backed by top tier investors, including Bessemer Venture Partners, Leitmotif, NVP Capital, and Sarah Smith Fund. Haptica’s advisors include famed roboticist Rodney Brooks, Benchmark Capital co-founder Andy Rachleff, and haptics expert Allison Okamura.

We’re a small and mighty team of builders driven to unlock touch as the new frontier. The next 6 months for us are all about building high quality hardware, quantifiably establishing the impact of tactile data on learning policies, and hiring A-players that can help us scale in 2027.

ABOUT THE ROLE

We’re looking for a Robot Learning Engineer to own the entire function at Haptica. We do not deploy robots, but we need to have in-house training capabilities to measure our effectiveness.

This includes creating our training pipeline, stress-testing our sensors, determining what data to collect and how, evaluating policies on the robot, and helping customers integrate our data into their stack. You’ll have extreme autonomy, but you won’t need to start from scratch. We have a Franka FR3 arm in the office, basic glove-based hand-tracking, and have laid the groundwork to train our own policies. We just need someone to take it further.

The sky is the limit for this role, but to start we have three broad goals:

1. Build our internal data collection pipeline alongside our hardware team

2. Drive and publish research to establish the role of tactile data in learning

3. Inform the roadmap for how our customers will use our data in their stack

WHAT YOU’LL DO

• Build our data collection, training and inference pipeline on off-the-shelf hardware integrated with Haptica.

• Train imitation learning policies (ACT, diffusion policy) with and without our tactile data on tasks where touch matters most: shear, occluded grasps, delicate objects.

• Define how we evaluate. Benchmark tasks that reuse publicly researched setups, metrics, and tooling to quantify the impact of our data and the variance of the impact across tasks.

• Extend data collection from teleop to egocentric glove data. Integrate hand tracking, test retargeting from a human hand to grippers and multi fingered hands, and run experiments that decide how many fingers we track.

• Influence hardware decisions based on what our sensing needs to deliver for learning. You'll have opinions on sampling rate, resolution, synchronization, drift, and calibration.

• Stay on the bleeding edge of how the world's best companies are running robot learning, and help us integrate Haptica into customers' data collection and teleoperation setup.

• Share your findings with the world through demo videos, writeups, and papers with our academic collaborators at Stanford, MIT, and more.
  

WHO YOU ARE

• You have a MS or PhD in robotics, CS, EE or a related field, or equivalent industry experience.

• Extensive experience in production-level software and ML engineering best practices.

• Experience with modern deep learning frameworks (e.g., PyTorch).

• You're an outstandingly strong written and verbal communicator skilled at simplifying complex topics to teammates, customers, investors, and friends. You make it look easy.

• You know the fundamentals for running imitation learning (diffusion policy, ACT or similar) from data collection to training, deployment, and debugging.

• You're comfortable with hardware. You've worked with cameras, IMUs, calibration, time synchronization across sensors. You have an intuition for hardware decisions that matter.

• You're comfortable with ambiguity, and get excited at the thought of building from scratch, defining your own constraints, and closely partnering with both internal and external teams.

BONUS IF

• You have experience with dexterous manipulation, multi-fingered hands, or retargeting

• You're passionate about tactile and have relevant research or experience in sensing

• You have experience taking egocentric or wearable data through to a trained policy

• You have a strong publication record in a relevant field

• Prior work on humanoids or highly dexterous robotic platforms