ML / Perception Engineer

Role Overview

Harkan Robotics is building autonomous robotic systems designed and engineered in Kerala for the global market. We're looking for an ML / Perception Engineer to own our robot's computer vision pipeline—from dataset engineering and model training to optimization and deployment on edge hardware. This role is ideal for someone who enjoys taking machine learning systems from experimentation to production and wants to see their work operating on real robots in real-world environments.

Responsibilities

• Own the end-to-end computer vision pipeline.
• Build, curate, and maintain training datasets.
• Train, evaluate, and optimize computer vision models.
• Convert and deploy models using ONNX and TensorRT.
• Optimize inference performance on NVIDIA Jetson device.
• Integrate perception systems with ROS 2-based robotic devices.
• Improve model accuracy and robustness under real-world operating conditions.
• Collaborate closely with navigation and robotics engineers.

Requirements

• 2+ years of production experience in Computer Vision.
• Strong experience with PyTorch, ONNX, and TensorRT (INT8 Optimization).
• Experience deploying ML models on NVIDIA Jetson devices.
• Experience with dataset engineering and annotation workflows.
• Experience integrating perception systems with ROS 2.
• Strong Python development skills.
• Experience debugging and optimizing real-world ML systems.

Skills

• PyTorch
• ROS 2
• Python
• TensorRT
• NVIDIA Jetson

Bonus Points

• Experience with segmentation models.
• RGB-D sensor fusion experience.
• Experience with Weights & Biases.
• Active learning workflows.
• Experience working on robotics or autonomous systems.
• Startup or product development experience.

What We Offer

• Direct access to founders.
• Real ownership from day one.
• Fast-paced startup environment.
• ESOP participation and long-term upside.
• Opportunity to build AI systems that operate on real robots.