Founding Engineer

Founding Robotics / ML Engineer
Location: Pittsburgh preferred; open to exceptional candidates elsewhere in the U.S. or Canada
Company: Momentedge
Stage: Early-stage / venture-backed
Visa Sponsorship: No
About Momentedge
Momentedge is building intelligence for understanding the behavior of deployed robots.
As robots move from structured automation toward increasingly complex and learned behaviors, understanding what actually happens in the field becomes harder. Engineers are left searching through large volumes of video, telemetry, trajectories, sensor data, and logs to find the small number of moments that matter.
We are building models that learn from multimodal robot data to detect and rank imperfect, unusual, or important moments of operation — helping robotics teams identify behavior worth investigating and data worth learning from.
Our long-term goal is to build an independent behavior intelligence layer for deployed robotics.
Momentedge is venture-backed and founded by a former Google product leader with experience across multimodal AI, ML infrastructure, cloud, and edge systems.
The Role
We are looking for a Founding Robotics / ML Engineer to help build the technical foundation of Momentedge.
You will work directly with the founder on questions such as:
How do we learn representations of normal robot behavior from largely unlabeled operational data?
How do we detect meaningful deviations without requiring engineers to define every failure mode in advance?
How should video, trajectories, robot state, telemetry, and other sensor signals be combined?
How do we distinguish merely unusual behavior from behavior actually worth an engineer's attention?
How can these models operate within the compute, storage, and connectivity constraints of deployed robots?
How do we evaluate behavior models when interesting events are rare and ground truth is limited?
This is not a role where the architecture has already been decided. You will help determine it.
What You'll Work On
You will:
Develop models for detecting and ranking meaningful robot behaviors and operational events.
Experiment with multimodal representations across video, telemetry, trajectories, robot state, and sensor data.
Build evaluation frameworks using real-world, lab, and simulated robot datasets.
Explore approaches including representation learning, temporal modeling, weak supervision, novelty/OOD detection, and multimodal learning.
Deploy and optimize inference for robotics and edge-compute environments.
Work with robotics researchers and industry teams to evaluate whether detected behaviors are genuinely useful.
Help shape the technical roadmap, architecture, and engineering culture of Momentedge.
As an early engineer, your scope will extend well beyond a narrowly defined ML role.
What We're Looking For
Must have
Strong Python and modern ML experience, ideally with PyTorch.
Hands-on experience developing or evaluating machine-learning models.
Experience with robotics, autonomous systems, computer vision, robot learning, or another ML domain involving physical systems.
Strong experimental instincts: you know how to determine whether a model is actually working rather than simply improving a benchmark.
Ability and desire to operate in an early-stage environment with significant technical ambiguity.
Willingness to build, test, discard assumptions, and iterate quickly.
Especially valuable
Robot learning, imitation learning, embodied AI, or VLA/VLM experience.
Multimodal or temporal representation learning.
Video understanding or computer vision.
ROS / ROS2.
Experience with trajectories, point clouds, LiDAR, or other robotic sensor modalities.
NVIDIA Jetson, CUDA, TensorRT, or other edge-inference experience.
Simulation environments such as MuJoCo, Isaac Sim, Gazebo, or similar.
Experience deploying ML systems on physical robots or autonomous machines.
Experience working with large, messy, weakly labeled datasets.
You do not need to check every box.
We care much more about technical depth, curiosity, and your ability to reason about robot behavior than about a particular degree or number of years of experience.
Who Might Be a Great Fit
You might be:
A robotics/ML engineer who wants substantially more ownership.
A researcher who wants to turn robot-learning ideas into a real product.
An engineer from autonomous vehicles, drones, warehouse robotics, manipulation, industrial robotics, or embodied AI.
An exceptional graduate student or recent graduate who has already built serious robotics/ML systems and wants a path toward a founding role.
Why Join Now
You'll join at the stage where fundamental technical decisions are still open.
You'll have the opportunity to help define how Momentedge models robot behavior, build the first generation of the technology, work directly with robotics researchers and companies, and grow into significant technical leadership as the company scales.
This is a high-ownership role for someone who wants to help build a robotics company from the beginning.
If you've worked on robot learning, perception, autonomy, multimodal ML, or related systems — and the problem of understanding what robots actually do in the real world sounds interesting — we'd like to hear from you.

Work arrangement
Yes