2+ Graphics Processing Unit (GPU) Jobs in Mumbai | Graphics Processing Unit (GPU) Job openings in Mumbai

We’re looking for an experienced AI Developer to join our team and drive the design, development, and deployment of advanced computer-vision and LLM-based solutions. You’ll build real-time object detection & tracking pipelines (LLM, YOLOv8/11), develop face-recognition models on video feeds with LLM support, and manage GPU-powered server infrastructure end-to-end.

Key Responsibilities

• Model Development & Optimization
• Design, train, and fine-tune LLMs, YOLOv8/11 models for multi-class object detection.
• Implement face-recognition pipelines leveraging state-of-the-art LLMs and embedding techniques.
• Optimize inference speed and accuracy for real-time video processing on GPUs.
• Video Analytics & Tracking
• Develop multi-camera object-tracking algorithms across disparate video sources.
• Integrate OpenCV, DeepSORT (or similar), and custom tracking logic for robust performance.
• Infrastructure & Deployment
• Provision and maintain Linux servers with SSH, Docker, and GPU drivers (NVIDIA CUDA/cuDNN).
• Automate CI/CD pipelines for model training, evaluation, and deployment.
• Monitor GPU utilization, troubleshoot performance bottlenecks, and ensure high availability.
• Collaboration & Documentation
• Work closely with product managers, data engineers, and front-end teams to integrate AI services via REST/gRPC APIs.
• Write clear technical documentation, API specs, and best-practice guides.

Required Qualifications

• Bachelor’s or Master’s in Computer Science, Electrical Engineering, or related field.
• 2-3+ years hands-on experience with:
• YOLOv8 or YOLOv11 (training, inference, transfer learning)
• Face recognition frameworks (InsightFace, FaceNet, ArcFace) and LLM embeddings (e.g., Hugging Face Transformers)
• Python ecosystem: PyTorch/TensorFlow, OpenCV, NumPy, scikit-learn
• Strong Linux skills: server provisioning, SSH, shell scripting.
• Proficiency in managing GPU servers (NVIDIA CUDA, Docker GPU containers)
• Experience deploying models in production (AWS/GCP/Azure or on-prem clusters).

Preferred Qualifications

• Familiarity with video-streaming protocols (RTSP/WebRTC).
• Knowledge of microservices architecture and API gateways.
• Exposure to MLOps platforms (MLflow, Kubeflow).
• Strong debugging, profiling, and performance-tuning abilities.

JioTesseract, a digital arm of Reliance Industries, is India's leading and largest AR/VR organization with the mission to democratize mixed reality for India and the world. We make products at the cross of hardware, software, content and services with focus on making India the leader in spatial computing. We specialize in creating solutions in AR, VR and AI, with some of our notable products such as JioGlass, JioDive, 360 Streaming, Metaverse, AR/VR headsets for consumers and enterprise space.

Mon-fri role, In office, with excellent perks and benefits!

Position Overview

We are seeking a Software Architect to lead the design and development of high-performance robotics and AI software stacks utilizing NVIDIA technologies. This role will focus on defining scalable, modular, and efficient architectures for robot perception, planning, simulation, and embedded AI applications. You will collaborate with cross-functional teams to build next-generation autonomous systems 9

Key Responsibilities:

1. System Architecture & Design

● Define scalable software architectures for robotics perception, navigation, and AI-driven decision-making.

● Design modular and reusable frameworks that leverage NVIDIA’s Jetson, Isaac ROS, Omniverse, and CUDA ecosystems.

● Establish best practices for real-time computing, GPU acceleration, and edge AI inference.

2. Perception & AI Integration

● Architect sensor fusion pipelines using LIDAR, cameras, IMUs, and radar with DeepStream, TensorRT, and ROS2.

● Optimize computer vision, SLAM, and deep learning models for edge deployment on Jetson Orin and Xavier.

● Ensure efficient GPU-accelerated AI inference for real-time robotics applications.

3. Embedded & Real-Time Systems

● Design high-performance embedded software stacks for real-time robotic control and autonomy.

● Utilize NVIDIA CUDA, cuDNN, and TensorRT to accelerate AI model execution on Jetson platforms.

● Develop robust middleware frameworks to support real-time robotics applications in ROS2 and Isaac SDK.

4. Robotics Simulation & Digital Twins

● Define architectures for robotic simulation environments using NVIDIA Isaac Sim & Omniverse.

● Leverage synthetic data generation (Omniverse Replicator) for training AI models.

● Optimize sim-to-real transfer learning for AI-driven robotic behaviors.

5. Navigation & Motion Planning

● Architect GPU-accelerated motion planning and SLAM pipelines for autonomous robots.

● Optimize path planning, localization, and multi-agent coordination using Isaac ROS Navigation.

● Implement reinforcement learning-based policies using Isaac Gym.

6. Performance Optimization & Scalability

● Ensure low-latency AI inference and real-time execution of robotics applications.

● Optimize CUDA kernels and parallel processing pipelines for NVIDIA hardware.

● Develop benchmarking and profiling tools to measure software performance on edge AI devices.

Required Qualifications:

● Master’s or Ph.D. in Computer Science, Robotics, AI, or Embedded Systems.

● Extensive experience (7+ years) in software development, with at least 3-5 years focused on architecture and system design, especially for robotics or embedded systems.

● Expertise in CUDA, TensorRT, DeepStream, PyTorch, TensorFlow, and ROS2.

● Experience in NVIDIA Jetson platforms, Isaac SDK, and GPU-accelerated AI.

● Proficiency in programming languages such as C++, Python, or similar, with deep understanding of low-level and high-level design principles.

● Strong background in robotic perception, planning, and real-time control.

● Experience with cloud-edge AI deployment and scalable architectures.

Preferred Qualifications

● Hands-on experience with NVIDIA DRIVE, NVIDIA Omniverse, and Isaac Gym

● Knowledge of robot kinematics, control systems, and reinforcement learning

● Expertise in distributed computing, containerization (Docker), and cloud robotics

● Familiarity with automotive, industrial automation, or warehouse robotics

● Experience designing architectures for autonomous systems or multi-robot systems.

● Familiarity with cloud-based solutions, edge computing, or distributed computing for robotics

● Experience with microservices or service-oriented architecture (SOA)

● Knowledge of machine learning and AI integration within robotic systems

● Knowledge of testing on edge devices with HIL and simulations (Isaac Sim, Gazebo, V-REP etc.)