Computer Vision Engineer I

Big Bang Boom Solutions Private Limited; Chennai District; On-site; INR 1,800,000 - 2,400,000; Full time

Job summary

Big Bang Boom Solutions Private Limited in Chennai seeks an experienced computer vision engineer to lead the development of spatial mapping, 3D perception, and visual navigation for autonomous drones. The role combines vision algorithms, deep learning, and real-time edge deployment.

The ideal candidate will drive CAM-based autonomy using CUDA, PyTorch/TensorFlow, ROS/ROS2, and Jetson platforms, with leadership responsibility across multiple CV domains.

Qualifications

• 6+ years in computer vision, image processing, and deep learning.
• Expert in object detection, tracking, segmentation for real-world applications.
• Strong foundation in 3D vision, SLAM, and camera geometry.
• Hands-on experience with CNN architectures and vision transformers.
• Expertise in CUDA programming and GPU optimization.
• Experience with TensorRT, ONNX for model deployment.
• Proficiency in embedded vision systems (Jetson platforms).

Responsibilities

• Lead computer vision teams for autonomous drone perception and navigation systems.
• Architect scalable vision pipelines for real-time aerial applications.
• Mentor engineers and drive technical excellence in vision-based autonomy.
• Implement small object detection optimized for aerial imagery.
• Build multi-object tracking (SORT, DeepSORT, ByteTrack) for dynamic environments.
• Deploy transformer-based models (DETR, Swin Transformer) for advanced perception.
• Create semantic segmentation for terrain classification and scene understanding.
• Implement Visual SLAM (ORB-SLAM3, RTAB-Map) for localization and mapping.
• Design depth estimation using stereo vision and monocular depth networks (MonoDepth, MiDaS).
• Build 3D reconstruction pipelines from aerial imagery.
• Develop point cloud processing using PCL and Open3D.
• Develop Visual-Inertial Odometry (VIO) systems for GPS-denied navigation.
• Implement optical flow for velocity estimation and obstacle avoidance.
• Design feature detection and matching (SIFT, ORB, SuperPoint).
• Build visual servoing for precision landing and target tracking.
• Optimize models for real-time inference (<50ms) on Jetson platforms.
• Implement CUDA, TensorRT, ONNX Runtime optimizations.
• Develop model quantization and pruning for edge deployment.
• Deploy with mixed precision inference (FP16, INT8).
• Expert proficiency in C++ and Python.
• Integrate vision systems with PX4/ArduPilot using ROS/ROS2.
• Implement camera calibration and multi-camera fusion.
• Manage timelines, deliverables, and technical documentation.
• Define performance benchmarks for accuracy, latency, and robustness.
• Coordinate with hardware teams on sensor integration.

Skills

6+ years in computer vision

Object detection & tracking

3D vision & SLAM

CUDA programming

Embedded vision (Jetson)

C++ / Python

ROS/ROS2

Education

B.E./B.Tech in Computer Science or ECE

M.Tech in Computer Vision, AI or related

Tools

OpenCV

PyTorch/TensorFlow

CUDA

ROS/ROS2

Docker

GIT

Job description

• B.E./B.Tech in Computer Science or Electronics & Communication Engineering (ECE)
• M.Tech in Computer Vision, AI, or related fields is preferred

Job Summary

We are seeking an experienced engineer with expertise in computer vision to lead the development of spatial mapping, 3D perception, and visual navigation systems for autonomous drones. The ideal candidate will combine deep knowledge in vision algorithms, deep learning, and real-time edge deployment.

Key Responsibilities

Technical Leadership

• Lead computer vision teams for autonomous drone perception and navigation systems
• Architect scalable vision pipelines for real-time aerial applications
• Mentor engineers and drive technical excellence in vision-based autonomy
• Implement small object detection optimized for aerial imagery
• Build multi-object tracking (SORT, DeepSORT, ByteTrack) for dynamic environments
• Deploy transformer-based models (DETR, Swin Transformer) for advanced perception
• Create semantic segmentation for terrain classification and scene understanding

3D Vision & Spatial Mapping

• Implement Visual SLAM (ORB-SLAM3, RTAB-Map) for localization and mapping
• Design depth estimation using stereo vision and monocular depth networks (MonoDepth, MiDaS)
• Build 3D reconstruction pipelines from aerial imagery
• Develop point cloud processing using PCL and Open3D

Visual Navigation

• Develop Visual-Inertial Odometry (VIO) systems for GPS-denied navigation
• Implement optical flow for velocity estimation and obstacle avoidance
• Design feature detection and matching (SIFT, ORB, SuperPoint)
• Build visual servoing for precision landing and target tracking

Edge AI & GPU Optimization

• Optimize models for real-time inference (<50ms) on Jetson platforms
• Implement CUDA, TensorRT, ONNX Runtime optimizations
• Develop model quantization and pruning for edge deployment
• Deploy with mixed precision inference (FP16, INT8)

Programming & System Integration

• Expert proficiency in C++ and Python
• Strong experience with OpenCV, PyTorch/TensorFlow, CUDA
• Integrate vision systems with PX4/ArduPilot using ROS/ROS2
• Implement camera calibration and multi-camera fusion

Project Management

• Manage timelines, deliverables, and technical documentation
• Define performance benchmarks for accuracy, latency, and robustness
• Coordinate with hardware teams on sensor integration

Required Skills & Qualifications

Core Competencies:

• 6+ years in computer vision, image processing, and deep learning
• Expert in object detection, tracking, segmentation for real-world applications
• Strong foundation in 3D vision, SLAM, and camera geometry
• Hands-on experience with CNN architectures and vision transformers
• Expertise in CUDA programming and GPU optimization
• Experience with TensorRT, ONNX for model deployment
• Proficiency in embedded vision systems (Jetson platforms)

Drone-Specific:

• Understanding of aerial imaging challenges (altitude, motion blur, perspective)
• Knowledge of PX4/ArduPilot and MAVLink protocol
• Experience with gimbal stabilization and camera systems

Software Engineering:

• Strong data structures and algorithms foundation
• Experience with ROS/ROS2, Git, Docker, CI/CD
• Proven leadership and team management skills

Preferred Qualifications

• Experience with VIO and sensor fusion (Kalman filters, EKF/UKF)
• Knowledge of reinforcement learning for vision-based control
• Publications in CVPR, ICCV, ECCV, ICRA, IROS
• Open-source contributions (OpenCV, ROS)
• Experience with photogrammetry and 3D modeling
• Familiarity with MLOps (MLflow, TensorBoard)
• Knowledge of simulation platforms (Gazebo, AirSim)
• Experience with thermal/multispectral imaging