Machine Learning Engineer

Kinetic Automation; Oakland (CA); On-site; USD 120,000 - 180,000; Full time

Benefits offered by this job

Competitive salary and equity package

Comprehensive health and dental保险

Job summary

Kinetic is building robotic repair centers powered by proprietary software and AI. We are hiring for a production-minded ML role to develop, evaluate, and deploy vision models for semantic/instance segmentation and object/damage detection across 2D and 3D data.

You will work with a small,cross-functional team in Orange County/Oakland. Ideal candidates have deep-learning experience in PyTorch, transformers for vision, and 3D data, plus strong Python skills and production tooling (Docker,

Qualifications

• BS or MS in CS/EE/Math or related field with deep-learning coursework or projects.
• Hands-on experience training and evaluating deep models for segmentation/detection in PyTorch, with data prep, augmentations, losses, and metrics (IoU/AP).
• Knowledge of transformers in vision (e.g., ViT, DETR) and self-attention concepts.
• Experience with 3D/depth data and multi-view geometry.
• Proficiency in Python and the PyTorch ecosystem (torchvision, Detectron2/MMDetection, Hugging Face Transformers).
• Experience with Python services (FastAPI/Flask) and Docker; AWS (S3/Batch/EC2/ECR) is a plus.
• Strong communication skills for tidy PRs, experiment logs, and design notes.

Responsibilities

• Collaborate on model development with training loops, losses, augmentations, and evaluations in PyTorch.
• Summarize papers/PRs and propose small, testable improvements.
• Contribute to datasets by labeling guidelines, splits, quality checks, and data versioning.
• Run experiments to track metrics, perform ablations, and document findings.
• Provide production support by exporting models, writing inference code, adding tests, and profiling performance.
• Work cross-functionally with backend engineers on APIs, containers, CI, and with labeling teams on edge cases.

Skills

PyTorch segmentation/detection

Transformers in vision

3D vision / depth data

Python programming

Model evaluation (IoU, AP)

Communication

Education

BS or MS in CS/EE/Math

Tools

Docker

AWS (S3/Batch/EC2)

FastAPI/Flask

MMDetection/Segmentation

Detectron2

Hugging Face Transformers

PyTorch

Job description

About Kinetic

Kinetic Automation is building a network of automated repair centers for modern vehicles. The auto industry is transitioning from mechanically complex vehicles to mechanically simple ones with complex software and technology. Kinetic aims to be the primary infrastructure-as-a-service for servicing future vehicles with our robotic repair centers, powered by our proprietary software and AI. We are a strong team of experienced robotics + automotive + shared mobility enthusiasts who have worked in self-driving, mapping, lidar, motorsport, ride‑sharing and ghost kitchens. We are a venture backed startup (Series B) with a clear go‑to‑market strategy and meaningful revenue.

About the role

You will be a part of a small, production‑minded ML team based in Orange County/Oakland. You’ll collaborate with other engineers and researchers to develop, evaluate, and help deploy vision models for tasks like semantic/instance segmentation and object/damage detection across 2D and 3D data.

Experience & Skills Required

• BS or MS in CS, EE, Math, or a related field with solid deep‑learning coursework or projects (e.g., grad‑level ML/CV/DL or equivalent independent work).
• Hands‑on experience training and evaluating deep models for segmentation and detection in PyTorch (or similar), including data prep, augmentations, losses, and metrics (IoU/AP).
• Working knowledge of transformer and LLM building blocks applied to vision, including self‑attention, positional encodings, tokenization, and mapping these ideas to vision models (e.g., ViT, DETR, Mask2Former).
• Practical exposure to 3D/depth data, including familiarity with point clouds, camera geometry (intrinsics/extrinsics), basic calibration, and multi‑view geometry.
• Proficiency in Python and the relevant tech stack: PyTorch, torchvision, Detectron2 or MMDetection/Segmentation, and Hugging Face Transformers.
• Experience with Python services (FastAPI/Flask), Docker, and AWS services (S3, Batch/EC2, ECR) is preferred.
• Strong communication skills with the ability to write tidy PRs, experiment logs, and short design notes to ensure reproducibility.

Responsibilities

• Collaborate on model development by implementing training loops, losses, augmentations, and evaluations using PyTorch.
• Keep current with the industry by summarizing relevant papers and PRs, and proposing small, testable improvements.
• Contribute to datasets by helping define labeling guidelines, curating splits, running quality checks, and maintaining data versioning.
• Run experiments to track metrics, perform ablations, write clear experiment notes, and present findings.
• Provide production support by exporting models, writing basic inference code, adding tests, and assisting with performance profiling.
• Work cross‑functionally, partnering with backend engineers on APIs, containers, and CI, and with ops/labeling teams on edge cases and feedback loops.

Benefits

• Competitive salary and equity package
• Comprehensive health and dental insurance
• Retirement savings plan.
• Paid time off and holidays

Kinetic is an equal opportunity employer.

We celebrate diversity and are committed to creating an inclusive environment for all employees.

We do not discriminate based on race, color, religion, gender, gender expression, age, national origin, disability, marital status, sexual orientation, military status, or any protected attribute.

We encourage qualified candidates from all backgrounds to apply and join us in our mission.

If you require accommodation at any stage of the application process due to a disability, please let us know.