Robotics Developer – Perception and AI

Ottonomy; India; On-site; INR 1,500,000 - 3,500,000; Full time

Job summary

Ottonomy is seeking a Robotics Developer to advance camera-based perception for autonomous robots operating in dynamic environments. You will deploy perception pipelines, integrate diverse camera systems, and optimize models for edge devices.

The role emphasizes moving prototypes to reliable field deployments, applying deep learning for object detection, scene understanding, and multimodal reasoning. Collaboration across robotics and hardware teams is essential.

Qualifications

• Bachelor's or Master's degree in Computer Science, Robotics, Electronics, Electrical Engineering, AI, or related field.
• 2+ years of professional experience in robotics, computer vision, autonomous systems, or applied AI.
• Strong programming skills in Python and C++.
• Experience deploying and debugging camera systems on real hardware.
• Strong understanding of image processing, camera calibration, coordinate transformations, and computer-vision fundamentals.
• Hands-on experience with deep-learning frameworks such as PyTorch or TensorFlow.
• Experience with ROS or ROS 2.

Responsibilities

• Develop, deploy, and maintain camera-based perception systems on autonomous robots.
• Integrate monocular, stereo, depth, USB, MIPI, GMSL, or similar camera systems.
• Build and optimize camera pipelines using tools such as GStreamer, OpenCV, ROS, and NVIDIA multimedia frameworks.
• Diagnose camera, driver, synchronization, latency, frame-drop, and hardware-interface issues.
• Develop deep-learning models for object detection, segmentation, tracking, depth estimation, scene understanding, and anomaly detection.
• Train, fine-tune, evaluate, and deploy neural-network models on edge-compute platforms.
• Work with Vision-Language Models for scene interpretation, semantic reasoning, operator assistance, and robot-task understanding.
• Explore and integrate Vision-Language-Action models and multimodal policies for robotic applications.
• Build datasets, annotation pipelines, evaluation frameworks, and model-monitoring tools.
• Optimize models using TensorRT, ONNX, CUDA, quantization, pruning, or other acceleration techniques.
• Integrate perception and AI components with ROS-based navigation and control systems.
• Profile and optimize CPU, GPU, memory, latency, and power consumption on embedded hardware.
• Conduct testing on physical robots and support field deployments.
• Collaborate with robotics, controls, navigation, hardware, cloud, and operations teams.
• Document system architecture, deployment procedures, experiments, and performance results.

Skills

Python

C++

ROS / ROS 2

OpenCV

Deep learning

GStreamer

CUDA

TensorRT

Education

Bachelor's or Master's degree in Computer Science, Robotics, Electronics, Electrical Engineering, AI

Tools

PyTorch

TensorFlow

ONNX

GStreamer

OpenCV

NVIDIA Jetson

Job description

About the Role

We are looking for a Robotics Developer with strong experience in camera-based perception, deep learning, and modern vision-language models. You will work on real-world autonomous robots operating in dynamic environments. The role involves deploying and optimizing camera systems, building perception pipelines, integrating deep-learning models, and exploring Vision‑Language Models (VLMs) and Vision‑Language‑Action models (VLAs) for robotic decision‑making and interaction. This is a hands‑on engineering role requiring experience moving systems from prototypes and research environments to reliable deployment on physical robots.

Key Responsibilities

• Develop, deploy, and maintain camera‑based perception systems on autonomous robots.
• Integrate monocular, stereo, depth, USB, MIPI, GMSL, or similar camera systems.
• Build and optimize camera pipelines using tools such as GStreamer, OpenCV, ROS, and NVIDIA multimedia frameworks.
• Diagnose camera, driver, synchronization, latency, frame‑drop, and hardware‑interface issues.
• Develop deep‑learning models for object detection, segmentation, tracking, depth estimation, scene understanding, and anomaly detection.
• Train, fine‑tune, evaluate, and deploy neural‑network models on edge‑compute platforms.
• Work with Vision‑Language Models for scene interpretation, semantic reasoning, operator assistance, and robot‑task understanding.
• Explore and integrate Vision‑Language‑Action models and multimodal policies for robotic applications.
• Build datasets, annotation pipelines, evaluation frameworks, and model‑monitoring tools.
• Optimize models using TensorRT, ONNX, CUDA, quantization, pruning, or other acceleration techniques.
• Integrate perception and AI components with ROS‑based navigation and control systems.
• Profile and optimize CPU, GPU, memory, latency, and power consumption on embedded hardware.
• Conduct testing on physical robots and support field deployments.
• Collaborate with robotics, controls, navigation, hardware, cloud, and operations teams.
• Document system architecture, deployment procedures, experiments, and performance results.

Required Qualifications

• Bachelor’s or Master’s degree in Computer Science, Robotics, Electronics, Electrical Engineering, Artificial Intelligence, or a related field.
• 2 or more years of professional experience in robotics, computer vision, autonomous systems, or applied AI.
• Strong programming skills in Python and C++.
• Experience deploying and debugging camera systems on real hardware.
• Strong understanding of image processing, camera calibration, coordinate transformations, and computer‑vision fundamentals.
• Hands‑on experience with deep‑learning frameworks such as PyTorch or TensorFlow.
• Experience with computer‑vision models, including object detection, segmentation, classification, or tracking.
• Experience with ROS or ROS 2.
• Familiarity with Linux development and debugging.
• Experience deploying models on NVIDIA GPUs or edge‑compute platforms.
• Understanding of model training, validation, benchmarking, and dataset management.
• Strong debugging and problem‑solving skills.

Preferred Qualifications

• Experience with Vision‑Language Models such as CLIP, LLaVA, Qwen‑VL, InternVL, Florence, or similar multimodal models.
• Experience with Vision‑Language‑Action models, imitation learning, behaviour cloning, reinforcement learning, or robot foundation models.
• Experience fine‑tuning multimodal models using LoRA, QLoRA, instruction tuning, or related techniques.
• Experience with NVIDIA Jetson platforms, CUDA, TensorRT, DeepStream, or JetPack.
• Experience with GStreamer and hardware‑accelerated video pipelines.
• Experience with MIPI CSI, GMSL, FPD‑Link, RealSense, stereo, or depth cameras.
• Experience with multimodal sensor fusion involving cameras, LiDAR, radar, IMU, or odometry.
• Familiarity with transformer architectures and large‑language‑model inference.
• Experience with Docker, Git, CI/CD, and cloud‑based model‑training workflows.
• Experience deploying autonomous robots in outdoor, industrial, warehouse, airport, campus, or public environments.
• Exposure to SLAM, localization, navigation, motion planning, or robot‑control systems.

Technical Skills

• Languages: Python, C++
• Robotics: ROS, ROS 2, TF, sensor integration
• Computer Vision: OpenCV, camera calibration, detection, segmentation, tracking
• Deep Learning: PyTorch, TensorFlow, transformers
• Multimodal AI: VLMs, VLAs, vision transformers, language‑conditioned policies
• Deployment: ONNX, TensorRT, CUDA, quantization
• Video Systems: GStreamer, DeepStream, V4L2
• Platforms: NVIDIA Jetson, embedded Linux, GPU‑based edge systems
• Tools: Docker, Git, Linux debugging and profiling tools

What We Value

• A strong hands‑on engineering mindset.
• Ability to debug across software, operating systems, drivers, cameras, and hardware.
• Interest in deploying AI on real robots rather than working only with offline datasets.
• Ability to balance model accuracy with latency, compute, reliability, and operational constraints.
• Ownership of projects from experimentation through field deployment.
• Comfort working in a fast‑moving robotics environment where systems are tested under real‑world conditions.

What You Will Work On

• Camera‑based perception for autonomous mobile robots.
• Small‑object and obstacle detection in complex environments.
• Multimodal scene understanding using VLMs.
• Language‑conditioned robot behaviours and VLA‑based research.
• Edge deployment and optimization of deep‑learning models.
• Reliable perception pipelines for continuously operating robots.
• Tools for data collection, model evaluation, diagnostics, and field debugging.